Energy-Balanced Dynamic Source Routing Protocol for Wireless Sensor Network Kok-Poh Ng, Charalampos Tsimenidis School of Electrical and Electronic Engineering Newcastle University Newcastle upon Tyne, NE1 7RU, UK k.p.ng@ncl.ac.uk , charalampos.tsimenidis@ncl.ac.uk Abstract—With the increasing deployment of unmanned and energy-constrained sensor devices in large-scale wireless sensor networks, energy efficiency and network lifetime have become key considerations in designing ad hoc routing protocols. In this work, we propose a fully distributed, multi-path load-balancing routing protocol based on Dynamic Source Routing (DSR) to improve network lifetime performance. The new protocol is simulated in ns-2 and compared with the commonly used Destination-Sequenced Distance Vector (DSDV), Ad hoc On- Demand Distance Vector (AODV) and DSR protocols. Our simulation results show that the new routing protocol has improved network lifetime significantly without sacrificing packet delivery performance. Keywords- wireless sensor network; load balancing; network lifetime; energy-aware; DSDV; AODV; DSR I. INTRODUCTION In recent years, wireless sensor network (WSN) has found its applications in the areas of Environment Monitoring and Forecasting, Health and Medical Care, Underwater Communications, Smart Energy, and Building and Home Automation industries. WSN operates in ad hoc mode, in which wireless clients (nodes) have an additional role of forwarding data from other clients towards their final destinations in a multi-hop operation, i.e. performing a routing function. The sensor nodes form a routing network as and when needed and the network topology could change rapidly from time to time due to the mobility of the sensor nodes. The operational environment for these sensor nodes poses additional constraints and challenges to the researchers and developers. As sensor nodes are typically deployed in large quantities in unmanned areas, and in most cases are non- retrievable, the cost factor and operational environment limit the physical form factor, processing capability, memory size, and battery power of these sensor nodes. Such challenges have generated great interests within the research community and many new algorithms and protocols have been proposed. With the trend of integrating different sensing applications towards the Internet of Things (IOT)[1], heterogeneous WSN with sensors of different capabilities will become a common scene. In heterogeneous WSN, it is important for the routing protocol to differentiate and balance the load among these different sensor nodes to achieve the optimal outcome. In this paper, we address the problem of network lifetime for energy-constrained sensor networks and propose Energy- Balanced Dynamic Source Routing (EB-DSR) protocol based on Dynamic Source Routing (DSR) [2]. The remainder of this paper is organized as follows: In section II, a brief description of related work in ad hoc routing protocols is provided. In section III, we propose a new energy- balanced routing algorithm. Section IV describes the network model and the performance metrics used in the simulations. Section V presents the simulation results and analysis; and finally we draw our conclusions in section VI. II. RELATED WORK A. Energy-aware Routing Most commonly used ad hoc routing protocols such as Destination-Sequenced Distance Vector (DSDV), Ad hoc On- Demand Distance Vector (AODV) and DSR use hop-count metric for route selection, which does not take into account of energy cost for utilising a particular link. As many ad hoc devices are energy-constrained, energy consumption has become a key consideration when evaluating routing protocols. Proposals to improve energy performance for ad hoc routing protocols have been made by incorporating energy cost for individual links. A metric based on minimum transmission energy is proposed in [3]; In [4], an effective total transmission energy metric that takes into consideration of link reliability and retransmissions is presented; and a more accurate energy model that accounts for total energy consumption of data packets, control packets and retransmissions is described in [5]. Energy efficiency in these variations of energy-aware routing protocols has improved. However, for a network with stationary nodes, these routing protocols almost always select the same path, the lowest cost path, given the same source- destination pair. The static behaviour of the path selection means that wireless nodes along the selected path have to work harder and consume more energy than those that are not on the selected path. The energies of these overworked nodes will naturally be depleted much faster than the other nodes. When this happens, there is a high possibility that the network will be partitioned, making some destination nodes unreachable. In WSN applications, even if the network is not partitioned, the