Optimal Rate Allocation of Compressed Data Streams in Multihop Sensor Networks

摘要

Consider that a transform coder is installed at each sensor node. In this paper, an analytical framework for optimal rate allocation of compressed data streams in wireless multihop sensor networks is presented. Compact, necessary and sufficient closed-form solution to the optimal rate allocation problem is derived based on the lower rate approximation of rate distortion function. Extensive simulations were conducted using real-world data traces, LEM and Intel Lab datasets. The simulation results show that the information of compression ratios of sensor data can be utilized to optimally distributing the limited transmission rate among sensors to greatly reduce the amount of transmitted data. Moreover, the performance gain of the optimal rate allocation increases exponentially at lower rates, lower compression ratios of sensor data and larger variation between compression ratios.