Mobility handling in MAC for wireless ad hoc networks

Wireless Sensor Networks(WSN) are becoming pop­ ular in a wide variety of applications. In a WSN, sensor nodes can be static or mobile, depending on applications. Handling mobility in a sensor network can pose interesting challenges in sensor protocol design, and special algorithms are needed to adapt to mobility in the network. In this article, we survey the current state of art in handling mobility in sensor networks. We first study how mobility affects the design of different layers in the sensor network architecture, and then take a closer look at mobility handling in the MAC layer. We also describe the results of simulations that we conducted to compare and analyze existing mobility-aware MAC protocols. Finally, we discuss open issues in dealing with mobility in sensor networks. I. IN T RO D U C T I O N A wireless sensor network (WSN) is an ad-hoc network of autonomous low-powered sensors that are spatially distributed and communicate wirelessly to cooperatively achieve a task. The past few years have seen these wireless sensors being used for diverse applications like environmental monitoring, battlefield surveillance, and traffic control. Fig. 1 shows a typical deployment of sensor nodes in data gathering applications. A sensor node consists of the following components: a microcontroller, a sensor, a transceiver, and a power source. The data gathered by the sensor devices is processed by the microcontroller and sent to a gateway for further analysis using the wireless transceiver. The power source for a sensor is typically a battery. These sensors are usually deployed to monitor physical parameters in harsh environments. Once deployed, it should operate without hu­ man intervention. As recharging batteries is difficult under such deployments, energy is the scarcest resource in sensor networks [1]. Besides being energy constrained, nodes in sensor networks can be mobile, due to several reasons. After deployment, nodes can fail due to hardware failure or running out of battery, causing the nodes to leave the network. New nodes need be added to the network to replace these failed nodes. In [2], such movements are classified as weak mobility. In contrast, strong mobility is characterized by sensor nodes physically moving through the network, because of the movement of the medium (wind, water) or the nodes themselves. With the advancement in sensor technologies, strong mo­ bility plays an important part in an increasing number of applications. In patient monitoring applications, sensor nodes are attached on patients and follow their mobility pattern as they walk around. In disaster recovery applications, workers