Wireless sensor node with low-power sensing

Abstract

Wireless sensor network consists of a large number of simple sensor nodes that collect information from external environment with sensors, then process the information, and communicate with other neighboring nodes in the network. Usually, sensor nodes operate with exhaustible batteries unattended. Since manual replacement or recharging of the batteries is not an easy, desirable or always possible task, the power consumption becomes a very important issue in the development of these networks. The total power consumption of a node is a result of all steps of the operation: sensing, data processing and radio transmission. In most published papers in literature it is assumed that the sensing subsystem consumes significantly less energy than a radio block. However, this assumption does not apply in numerous applications, especially in the case when power consumption of the sensing activity is comparably bigger than that of a radio. In that context, in this work we focus on the impact of the sensing hardware on the total power consumption of a sensor node. Firstly, we describe the structure of the sensor node architecture, identify its key energy consumption sources, and introduce an energy model for the sensing subsystem as a building block of a node. Secondly, with the aim to reduce energy consumption we investigate joint effectiveness of two common power-saving techniques in a specific sensor node: duty-cycling and power-gating. Duty-cycling is effective at the system level. It is used for switching a node between active and sleep mode (with the duty-cycle factor of 1%, the reduction of in dynamic energy consumption is achieved). Power-gating is used at the circuit level with the goal to decrease the power loss due to the leakage current (in our design, the reduction of dynamic and static energy consumption of off-chip sensor elements as constituents of sensing hardware within a node of is achieved). Compared to a sensor node architecture in which both energy saving techniques are omitted, the conducted MATLsimulation results suggest that in total, thanks to involving duty-cycling and power-gating techniques, a three order of magnitude reduction for sensing activities in energy consumption can be achieved.