, pp. 345 – 349, 2006.
© Springer-Verlag Berlin Heidelberg 2006
Energy-Aware Routing Analysis in Wireless
Chow Kin Wah, Qing Li, and Weijia Jia
Department of Computer Science,
City University of Hong Kong,
Kowloon, Hong Kong
Abstract. Applications of sensor networks have become an emerging technology
which can monitor a specific area and collect environmental data around the
district. The energy of sensor nodes is tightly constrained so that there is a need to
control the power consumption in node operations such as transmission and
routing. In this paper, we carry out analysis on energy-aware routing in order to
minimize the path loss. Our goal is to prolong the network lifetime and ease the
management of sensor networks.
Sensor networks are highly energy constrained. The sensor nodes are expected to work
for a year or more using the power supplied from the on-board batteries. Since it is
impractical to replace the batteries on thousands of sensor nodes, we need an
appropriate solution to manage the energy consumed by sensor nodes so that the
lifetime of the network can be extended.
All the operations in a sensor node consume certain amount of energy. But the most
amount of energy is dissipated in the radio circuit, especially during transmission.
Therefore, when determining the next hop during routing, the sensor nodes should use
as much local information as possible and make less data exchange with neighbours in
order to conserve energy.
In this paper, we carry out analysis on energy-aware routing in order to minimize the
path loss. Our goal is to prolong the network lifetime  and ease the management of
sensor networks. The rest of our paper is organized as follows. In section 2, we classify
the basic terminologies to be used in the subsequent sections. In section 3, analysis on
energy-aware routing for minimizing the path loss is conducted in depth, including
such issues as zone-based management, layer-based management, and path loss with
the free space model. Finally, we conclude this paper and suggest some further research
issues in the last section.
H.T. Shen et al. (Eds.): APWeb Workshops 2006, LNCS 3842
346 C.K. Wah, Q. Li, and W. Jia
2 Energy-Aware Routing
2.1 Zone-Based Management
In a sensors network, we want to monitor a specific area and collect the required
environmental data. To achieve this goal efficiently, the target environment is often
divided into zones and only one sensor node is deployed in each zone. In R-DCS ,
there are three types of sensor mode: monitor, replica and normal. Each zone has one
monitor node for each event-type and at most one replica node for each event-type.
Zones are divided into Z squares with equal size and identified with zj : j = 1, …, Z.
When we define zones in this way, we assign the zone ID arbitrarily or in ascending
order. The zone ID cannot provide any geographical meaning and there is no
relationship between adjacent zones. Nodes cannot make use of zone information to aid
energy-aware routing. A rule is needed to classify the zones so that any zone can tell its
physical location and energy-related information through its ID.
2.2 Layer-Based Management
Based on the requirements, we change the zones into layers in which the power
required for the sensor nodes to transmit data directly to the access point is within the
same range. Nodes should route data towards the layer which is closer to the access
point (AP). Each node should maintain: (1) its direct PTx to AP, or if direct link is not
possible, the direct PTx to its upper layer; (2) its upper layer’s direct PTx to AP. As
shown in Fig.2, we use circular lines to denote the boundary of layers and there are
three layers. In the center there is the access point. The distance between the boundary
of layer 1 and the access point is r1. Now we have layer number L = 3 and layers lj: j =
1, …, L and radius rj: j = 1, …, L. A node mij belongs to layer j if
Fig. 1. Use layers instead of zones to divide the areas
Energy-Aware Routing Analysis in Wireless Sensors Network 347
where mij is the monitor node of event type i at layer j. In Fig.2, we have 4 event types
and each node is responsible for sensing a particular event type.
2.3 Path Loss with Free Space Model
We use the free space model  to analyze the differences between layers through
the following equation:
d dBWG dBWG dBWP dBWP
The last two terms are the path loss between the transmitter and receiver. Assume
sensor nodes are uniformly distributed around AP and r1=10, r2=20 and r3=30, the
distribution of nodes and path loss in each layer are shown in Table 1.
Table 1. Path loss of routing data via upper layer and transmitting directly to the AP
Directly to AP
Route via upper layer
/ 1 d
/ 1 d
1100.11 200.01 200.01
Over 50% of nodes are located in the outer-most layers. If they need to transmit data
directly to AP, the path loss is 29.54dBW which means 10 times of loss in absolute
power when compared with layer 1. Three times of distance can introduce 10 times of
path loss in this small area. The greater the path loss, the larger amount of energy we
need to supply to achieve the same SINR (Signal-to-Interference- and-Noise-Radio).
This addresses the importance of multi-hop routing in a wireless sensors network.
When we move to routing indirectly through upper layer nodes, we find that path
loss can be fixed in a small value if nodes in the lower layer can reach the nodes in
upper layer within a constant distance. According to , the overall rate of energy
dissipation is minimized when all the hop distances are equal to D/K where D is the
total distance traveled, and K is the number of relays plus one.
In Fig.3, the direct receiving area of the AP is depicted in the dashed circle. Nodes
outside this circle have to relay their data through the other nodes. The numbers on the
arrows indicate the path loss between two nodes. We define the average path loss from
node i to AP through node j as follows:
348 C.K. Wah, Q. Li, and W. Jia
if #relay n = 0
if #relay n > 0
We calculate the average path loss of two paths <C, D, E> and <C, J, E>, and find
that <C, J, E> has always a smaller value. But if we look at the average path loss of <C,
J, I> and <C, K, I>, the values are the same when it is calculated upon node I. When this
happens we need to check the variances along both paths, and the path with smaller
variance should be chosen. This is because the path loss along this path is more similar
to the D/K and less energy will be consumed.
Fig. 2. Path loss between different nodes
Table 2. Average path loss along <C, D, E>
and <C, J, E>
Path Average path loss
<C, D, E>
<C, J, E>
5 . 6
Table 3. Average path loss along <C, J, I>
and <C, K, I>
Path Average path loss
<C, J, I>
5 . 6
<C, K, I>
In this paper, we have discussed the layer-based approach to manage the sensors
network and the path loss in routing through nodes of different layers. Based on our
study, it is advantageous by using the average path energy with the aid of variance of
path loss to determine the path in energy-aware routing. In the future, we will not only
consider energy consumption but also plan to investigate routing based on the data
attributes and event types.
Energy-Aware Routing Analysis in Wireless Sensors Network 349
This work is supported by CityU SRG grant no. 7001709.
1. M. Bhardwaj, T. Garnett, and A. P. Chandrakasan, “Upper Bounds on the Lifetime of Sensor
Networks”, in Proc. of the International Conference on Communications(ICC’01), 2001.
2. A. Ghose, J. Grossklags, and J. Chuang, “Resilient Data-Centric Storage in Wireless Ad-Hoc
Sensor Networks”, in Proc. of the International Conference on Mobile Data
3. K. Lassonen, “Radio Propagation Modelling”, 2003.
4. T. Rappaport, “Wireless Communications: Principles & Practice”, Prentice-Hall, 1996.
5. S. Singh, M. Woo and C. Raghavendra, “Power-Aware Routing in Mobile Ad Hoc
Networks”, MOBICOM ’98, 1998.
6. A. Hac, “Wireless Sensor Network Designs”, Wiley, 2003.