|
Ever since the advent of wireless telecommunication technologies as well as low-cost, low-power and multifunctional sensor nodes (SNs), Wireless Sensor Networks (WSNs) have become the focal point of attention for many researchers. They support highly varied applications such as military, environment monitoring, agriculture, transportation control, disaster, fire fighting and home applications. Severe power, computation, bandwidth and storage constraints make them different from traditional wireless networks. SNs in WSN possess sensing, processing and wireless communication capabilities. They are generally left unattended, e.g., in hostile environments such as battlefields and volcanoes, which makes it sometimes impossible to re-charge or replace their batteries. This makes it crucial to optimally utilize resources such as energy, transmission media etc in WSNs to extend the lifetime of the network.
There are a lot of approaches to enhance the energy efficiency of a WSN. Some use proper signal processing techniques or a scheduled sleep/awake pattern at physical or MAC layer while others use proper hardware selection for sensors, as they differ in their power consumption in various working modes e.g. MCU active for MICA2 is 8.0mA while for Telos is 1.8mA. Yet others install suitable operating system like Tiny-OS to reduce the computational costs, thereby saving energy. Thus energy efficiency must be provided at all networking layers. The IEEE 802.15.4 standard is specifically meant to support long battery life time.
The most common technique for energy saving in WSN is judicious selection of a suitable routing protocol. These protocols implement various schemes such as multicasting (SPIN), anycasting (Gossiping), unicasting (Directed Diffusion), aggregation, compression, selection (TEEN) and various logical arrangements of nodes in the network in order to achieve energy efficiency. Among various logical arrangements such as ring, tree, tiers etc the most common is clustering.
In clustering the SNs are logically arranged into groups having their respective access point called cluster head (CH). The members of cluster transmit the sensed data to their CH in a single-hop (LEACH) or multihop (PEGASIS) fashion. This data is aggregated at CH and is then transmitted to base station (BS), again in a single-hop or multihop manner. The key advantages achieved by clustering are energy efficiency, scalability and communicational efficiency.
Preventing multiple nodes to communicate simultaneously with the BS enhances the scalability and communication efficiency of the network. It also reduces collisions and thus enhances efficeincy at MAC layer.
|