SYSTEMS AND METHODS FOR OPTIMIZING COMPRESSION AND RECONSTRUCTION OF SENSOR DATA IN DISTRIBUTED COMPUTING NETWORKS

摘要

A distributed computing system and method of sensor data compression. The system includes sensor-enabled edge nodes arranged in a cluster. A plurality of transforms, each configured for compressing the raw data signal into a compressed data signal. A controller is configured to assign each of the edge nodes in the cluster with a different one of the transforms. The controller receives a compressibility metric from each of the edge nodes, each compressibility metric estimating a reconstruction accuracy of data compressed in accordance with the each of the different compression transforms assigned to each of the edge nodes. The controller selects a preferred transform based on the compressibility metric of the different transforms, and instructs the edge nodes to each use the preferred transform when compressing one or more subsequent data signals from the sensor.