Variable-Word-Length Coding for Energy-Aware Signal Transmission

This paper presents an energy-efficient method for data transmission through improved data compression. This is achieved by a new signal source coding method based on the Walsh transform. This research is motivated by the need for extended service life of wireless sensor networks, where individual sensor nodes are energy constrained. Theoretical background is introduced, and its effectiveness is established through simulations and experiments. The simulations are conducted using electrocardiogram (ECG) and bearing vibration signals, and the results are comparatively evaluated against the Haar transform coding, discrete cosine transform (DCT) coding, pulse code modulation, and fixed-word-length Walsh coding. Experimental verification is performed by implementing the algorithm on a wireless test bed. Results indicate that up to 58% energy reduction can be achieved for wireless transmission of ECG signals, as compared to other commonly used coding methods. For bearing vibration signal transmission, the energy consumed during transmission is comparable to that of the near-optimal DCT coding.

I. Introduction

Demands for timely and comprehensive gathering of information concerning environments and machines have been increasingly met by the widespread acceptance of wireless sensor networks. Wireless sensor networks enable portability and improved accessibility to locations where cable connection is limited. Given that wireless sensors are mostly powered by compact batteries with limited capacity, wireless sensor networks are generally constrained on operational life span. Recharging or battery replacement is typically not an option after the sensor nodes are deployed, and even if available, the service will be time consuming and costly to implement in many application scenarios.