Cell based energy density aware routing: a new protocol for improving the lifetime of wireless sensor networks
Introduction
Recent advances in wireless communication and microsensor technologies have made it possible to develop wireless networks of low-cost, small-sized sensor nodes [1]. These sensor nodes can be deployed in physical environments to collect information such as acoustic, light, and seismic data, in an autonomous manner [2]. This technology has potential applications in diverse areas, including surveillance, rescue efforts in disaster areas, and military intelligence collection.
An important feature of wireless sensor networks is that the sensor nodes are equipped with batteries that cannot be recharged. The sensor nodes have this feature because they are intrinsically immobile; typically they are embedded in physical structures or thrown into inhospitable terrain and then left unattended. Thus, wireless sensor networks differ from mobile ad-hoc networks (MANET), which usually have rechargeable and replaceable batteries. Given that battery failure leads to failure of the sensor node, which plays the dual role of data originator and data router [1], the depletion of node batteries in an area can potentially cause significant topological changes within the network due to broken connections. Such changes can lead to partitioning of the network into completely disconnected areas. In addition, if a subarea has no live sensor node due to battery depletion, creating a so-called void subarea, users cannot access information on the subarea and paths to users that pass through the subarea may be cut off. Sensor network users wish to keep receiving information from all subareas for the longest possible time.
Energy consumption by sensor nodes can be divided into three domains: sensing, communication, and processing. Here, we concentrate on energy consumption for communication because this domain is known to consume the greatest amount of power in modern processors. For example, Berkeley motes [23] consume 8 mJ to transmit and 4 mJ to receive a byte, whereas the CPU requires about 0.8 mJ to execute 208 cycles (roughly 100 instructions). Thus, an energy-aware routing protocol that prevents energy consumption from concentrating in a few subareas (i.e. fair energy consumption) is required so that network connectivity is enhanced and access to some areas is not lost. Energy-aware routing protocols [3], [4] for MANETs find a single optimal path for low energy consumption. In sensor networks, sensor nodes usually report data repeatedly to one or a few monitoring stations, known as sinks, along the path. Nodes along the path thus tend to consume more energy than other, idle nodes. Moreover, if multiple paths are used, nodes at the points of overlap among the paths will more quickly run out of energy, which may eventually lead to the creation of disconnected subnets. Another problem of existing power- or energy-aware routing protocols is that they do not take into account the delay from a source to a sink. If wireless sensor networks are constructed densely with a large number of nodes, packets can traverse the network in many short hops in order to save energy. However, such routing protocols entail unfeasibly long delays, and the addition of new sensor nodes to complement an existing sensor field will cause these delays to increase.
To tackle the problems outlined above, we propose the cell based energy density aware routing (CEDA), a routing protocol that uses energy density as a routing metric and is tailored to the unique features of wireless sensor networks. When the sensor field is divided into small subareas called cells, the energy density is defined as the weighted sum of the energies of all nodes in the cell. In each cell, a sensor node with a higher level of residual energy relays packets on behalf of other nodes in the same cell. This sensor node also forwards packets to the neighbor cell with the greatest residual energy density. This protocol ensures that energy is consumed more fairly by the sensor nodes, and thus that the sensor network runs longer. The lifetime of a cell is defined as the time required for all nodes in the cell to lose all of their energy. Increasing the cell lifetime decreases the frequency with which void cells appear, and thus helps to maintain the connectivity of the network and the ability to supervise all subareas. Because sensor nodes are intrinsically immobile and are used at densities several orders of magnitude higher than in MANETs, the CEDA is based on a cellular addressing scheme to group the sensor nodes. The cellular addressing approach has less overhead than a global addressing or position-based addressing [13] because it groups several nodes in the same area. Cellular addressing reduces communication burden due to sensor flooding. The CEDA routing protocol also guarantees the maximum hop count regardless of the node density of the network.
The remainder of this paper is organized as follows. Section 2 reviews related studies and compares other routing protocols with the CEDA protocol. The cellular addressing method and the basic service model of CEDA are introduced in Section 3. Section 4 provides a detailed description of the CEDA algorithms. Section 5 presents proof that CEDA guarantees the maximum hop count. A comparative performance evaluation using simulation is presented in Section 6. Section 7 concludes the paper.
Section snippets
Related work
Directed diffusion is a data-centric communication paradigm specifically designed for sensor networks [5]. Sinks use flooding to spread interests to the sensor network. Sensors matching an interest send their data to the sinks along multiple paths initially, and then gradually reinforce better paths. The use of multiple paths and the flooding of packets are likely to consume large amounts of energy, and do not take into account fair energy consumption. TTDD [20] is an overlay network protocol
Addressing
Traditional networks assign unique global addresses (e.g. IP addresses of the Internet) to nodes. Sensor nodes, however, may not have global identification addressing because of the large amount of overhead and the large number of nodes in a sensor network [1]. The user or monitoring station typically requires information from a specific local area. Thus, it is helpful that the nodes are identified using their geographic positions. In fact, this geographic information can be exploited in the
Cell based energy density aware routing algorithm
We now outline the CEDA algorithm, which consists of two phases, a route discovery phase and a reconstruction phase. During the route discovery phase, the sink disseminates a query including an attribute and its location. The sensor nodes matching the query send data packets to the location of the sink through a path chosen based on the energy density. During the reconstruction phase, routes can be changed to avoid uneven draining of the energy levels of the batteries in the various nodes.
In
Guaranteeing the maximum hop count
A query may include an update period that limits the maximum delay from the sensor to the sink. Some power- or energy-aware routing protocols in wireless ad-hoc networks use paths consisting of short hops to save energy. However, such protocols can potentially suffer from very long delays because they do not limit the maximum number of hops allowed. In the case of wireless sensor networks, which typically consist of many densely distributed nodes, this problem would become worse. In this
Evaluation
To test whether the CEDA method actually conserves energy and to ascertain how long each monitored area is available in a finite-energy sensor network under a given set of conditions, we carried out simulations of sensor networks under various conditions. CEDA was simulated in the ns-2.26 simulator [21]. The simulation model was a network consisting of 200 randomly distributed acoustic sensor nodes and two sinks in an area of 100 m×100 m. The sinks could move at low speed. The sensor were set and
Conclusions
This paper has described a new energy-aware routing protocol, cell based energy density aware routing (CEDA), which is specifically designed for wireless sensor networks. It uses energy density as a routing metric and finds forwarding paths using information on geographic locations. A key concept of CEDA is the creation of simple routes by grouping nodes within uniform subareas (or cells). Another key concept of the proposed protocol is avoiding the concentration of energy consumption in
References (23)
- et al.
A survey on sensor networks
IEEE Communication Magazine
(2002) - et al.
Maximum lifetime data gathering and aggregation in wireless sensor networks
(2002) - et al.
Power-aware routing in mobile ad hoc networks
(1998) - et al.
Minimum energy mobile wireless networks
IEEE Journal of Selected Areas in Communication
(1999) - et al.
Directed diffusion: a scalable and robust communication paradigm for sensor networks
(2000) - et al.
Energy-efficient communication protocol for wireless microsensor networks
(2000) - et al.
Data gathering algorithms in sensor networks using energy metrics
IEEE Transactions on Parallel and Distributed Systems
(2002) - et al.
Highly dynamic destination sequenced distance-vector routing (DSDV) for mobile computers
(1994) - et al.
Dynamic source routing in ad hoc wireless networks
(1996) - et al.
Ad hoc on demand distance vector routing
(1999)
A performance comparison of multi-hop wireless ad-hoc network routing protocols
Cited by (0)
- 1
Tel.: +82 1099814563; fax: +82 28788933.