Adaptive management schemes for mobile ad hoc networks

Mobile ad hoc networks (MANETs) are wireless communication networks which are of interest because of their flexibility and ease of deployment. MANET nodes are often powered by batteries, and their replacement is usually difficult. Inter-node transmission power thus constrains the network topology and affects the communication efficiency. Furthermore, the network connectivity changes continuously due to mobility. Hence, understanding node mobility and adaptive management of transmission power are essential for successful network operation. This work addresses the implementation of efficient power-aware MANET management schemes. First, we analyze mathematical models of node movement and propose a metric that quantifies mobility. Existing network control algorithms are usually evaluated using random mobility models. However, since such models employ incompatible mobility parameters, it is hard to compare the performance of different algorithms. It has been shown that the impact of mobility on the network performance is a function of route lifetime. We show that link duration has a nearly invariant relationship with route lifetime regardless of the adopted mobility model, and thus is a good mobility metric. Second, we investigate the issues of power control and link maintenance. Existing power control schemes are mainly intended for static or pseudo-static networks, and their effectiveness in highly mobile networks has not been demonstrated. We develop a novel algorithm, which adaptively controls transmission power, and reduces communication power needs by more than 50% compared to existing algorithms with homogeneous transmission range. We also analyze the impact of medium access control on network performance. We show that the widely used RTS/CTS handshake protocol may adversely affect the network throughput when communication power is adjusted to the minimum necessary level. We further present a means to maximize the network throughput. Third, we investigate the problem of optimally placing base station and relay nodes. Appropriate insertion of such nodes can reduce power consumption and improve network performance. We apply non-linear optimization techniques to node placement, and present distributed node placement techniques which place nodes among radio obstacles to minimize the energy consumption. Simulation results confirm that the efficiency of the proposed algorithms is comparable to that of an existing centralized algorithm.Ph.D.Computer Science & EngineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/63719/1/sungcho_1.pd