A multi-layered energy consumption model for smart wireless acoustic sensor networks

Gert Dekkers, Fernando Rosas, Steven Lauwereins, Sreeraj Rajendran, Sofie Pollin, Bart Vanrumste, Toon van Waterschoot, Marian Verhelst, Peter KarsmakersarXiv 2018 · arXiv

Introduction

Smart wireless acoustic sensor networks (WASNs) are increasingly deployed in smart city applications for tasks ranging from traffic monitoring to emergency detection. However, these "smart" microphones face a significant challenge: high power consumption. To optimize battery life, engineers must decide whether to process data locally or transmit raw audio. This paper introduces WASN-EM, a comprehensive energy consumption model designed to bridge the gap between machine learning development and hardware design.

The WASN-EM Framework

WASN-EM provides a holistic view of energy costs by breaking down consumption into three primary layers: sensing, processing, and communication. This modular approach allows researchers to predict the battery life of a sensor node based on specific hardware parameters and algorithmic complexity.

Overview of the multi-layered WASN-EM framework, illustrating the interaction between sensing, processing, and communication layers.

Layered Energy Modeling

The model's strength lies in its granular approach to each functional block:

  1. Sensing Layer: Accounts for the microphone, Low-Noise Amplifier (LNA), and Analog-to-Digital Converter (ADC) using physical parameters like the Noise Efficiency Factor to estimate current draw.
  2. Processing Layer: Moves beyond simple operation counts (FLOPs) to include the significant energy costs of memory access. It distinguishes between on-chip SRAM and off-chip DRAM, providing a more realistic energy profile for complex AI models.
  3. Communication Layer: Models the entire RF chain and uses statistical methods (geometric distributions) to account for re-transmissions caused by wireless fading or interference.
The processing layer model highlights the energy difference between on-chip and off-chip memory accesses alongside ALU operations.

Key Insights: The "Smart" Trade-off

A central finding of the WASN-EM research is that memory access is often the "hidden killer" of energy efficiency. By modeling data movement and in-place computation, the framework identifies the "break-even point" where the energy saved by local data compression outweighs the cost of the computation itself. Furthermore, the model's statistical approach to communication acknowledges that the cost of transmitting a bit is not fixed but depends on environmental factors and error rates.

Conclusion

WASN-EM offers a vital tool for the design of sustainable smart sensor networks. By integrating hardware-level details with high-level algorithmic requirements, it enables the development of energy-aware acoustic sensing systems. Future work will expand the model to include more complex interference scenarios and specialized hardware accelerators.

Summary updated

Abstract

Smart sensing is expected to become a pervasive technology in smart cities and environments of the near future. These services are improving their capabilities due to integrated devices shrinking in size while maintaining their computational power, which can run diverse Machine Learning algorithms and achieve high performance in various data-processing tasks. One attractive sensor modality to be used for smart sensing are acoustic sensors, which can convey highly informative data while keeping a moderate energy consumption. Unfortunately, the energy budget of current wireless sensor networks is usually not enough to support the requirements of standard microphones. Therefore, energy efficiency needs to be increased at all layers --- sensing, signal processing and communication --- in order to bring wireless smart acoustic sensors into the market. To help to attain this goal, this paper introduces WASN-EM: an energy consumption model for wireless acoustic sensors networks (WASN), whose aim is to aid in the development of novel techniques to increase the energy-efficient of smart wireless acoustic sensors. This model provides a first step of exploration prior to custom design of a smart wireless acoustic sensor, and also can be used to compare the energy consumption of different protocols.
BibTeX
@misc{dekkers2018multi,
  author = {Gert Dekkers and Fernando Rosas and Steven Lauwereins and Sreeraj Rajendran and Sofie Pollin and Bart Vanrumste and Toon van Waterschoot and Marian Verhelst and Peter Karsmakers},
  title = {{A multi-layered energy consumption model for smart wireless acoustic sensor networks}},
  year = {2018},
  eprint = {1812.06672},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1812.06672},
}
Figure