> 1 Combined Effects of Mobility, Congestion and Contention on Network Performance for IEEE 802.15.4 Based Networks Radosveta Sokullu, Cagdas Donertas Abstract — Besides being very application specific the design and performance evaluation of Wireless Sensor Networks (WSN) is also very dependent on another inherent feature of these networks - the mobility of the sensor nodes. A major challenge is to provide reliable and energy-efficient operation taking into consideration different mobility models. On the other hand, due to the multi-hop nature of many WSN local contention can lead to network-wide congestion and reduce both the efficiency and the lifetime of the network. In this paper we study the interdependence between end-to-end congestion and local contention taking into consideration different mobility scenarios. The work is based on the slotted CSMA/CA medium access control method adopted in IEEE 802.15.4 protocol specification for LR-WPAN and covers beacon- enabled mode. Random Waypoint Mobility Model (RWP) and Ad Hoc On Demand (AODV) routing protocol are used. Beacon-enabled mode is chosen because it includes regulating the activity period of each single node. The aim is to investigate the effects of changing local contention regulating parameters (activity periods, number of MAC layer retransmissions and transmission buffer size) on the overall network congestion in different mobility scenarios. Index Terms—IEEE 802.15.4 beacon-enabled mode, mobility in WSN, network performance. I. INTRODUCTION Wireless Sensor Networks (WSN) are intensively used in different applications in the military, health, environment, agricultural and smart office domain. WSN are very application specific and nodes are very resource restricted so in order to operate efficiently the WSN must be carefully designed to fit the specific user requirements. Efficient communication in sensor networks is an interesting research topic and is being explored by using both traditional (layered protocol stack) and non-traditional (cross-layer design) methods. Considering the dynamic behavior of the sensors, their very restricted resources as well as the challenges in the network conditions, cross-layer design is a good alternative to traditional layered approaches. The idea behind cross-layer is, sharing information among different protocol layers and increased interlayer interactions. In order to propose and implement effective cross layer solutions it is very important to have a clear picture of the existing interdependence between the layers and the key parameters that control it. In this respect one of the major issues is the interdependence between MAC and Transport layer operation as collisions and congestion are a major source of packet loss in WSN. Congestion causes non-efficient operation of WSN and is mainly due to contention caused by concurrent transmission, buffer overflows and the dynamically time varying conditions of the wireless channel. The WSN is a multi-hop network and localized congestion may easily diffuse to the whole network and strongly degrade the network performance. Congestion control is crucial in achieving reliable and energy efficient operation of the network. On the other hand mobility is and inherent feature of WSN and different scenarios pose different mobility models. One of the most widely used models is the Random Waypoint model. In this paper we investigate the relationship between local contention resolution and end-to-end congestion based on the Random Waypoint Mobility model and try to determine the effect of major parameters such as node density, transmission buffer size and operation mode on the network performance in different mobility scenarios. The work is guided by possible WSN application scenarios like for example a monitoring scenario where mobile agent nodes are deployed in a certain area to collect environmental specific data. The sink, which can also be a gateway, might be one of the nodes. Different nodes can take the role of a sink in a rotating manner in order to provide a more equal distribution of the energy consumption and prolong the lifetime of the network as a whole. Another scenario considered is the one when a fixed sink/gateway is collecting information from mobile agents distributed in a given area. A third scenario discussed is the case when a mobile sink moves randomly among static (but randomly deployed) nodes and collects information. The work is based on the IEEE 802.15.4 protocol for low-rate wireless personal area networks (LR-WPAN) which specifies a superframe structure allowing regulation of the node’s active/inactive period duration. Most of the related work so far has been concentrated on investigating the specifics of the IEEE 802.15.4 but usually a general scenario is considered. The authors of [25] develop extensions for the ns-2.31 simulator covering the IEEE 802.15.4 and conduct several sets of experiments to study its