Abstract

Recently, wireless ad hoc networks have witnessed a remarkable interest. In such networks, the medium spatial reuse has a tremendous impact on the network performance, since it determines the number of concurrent transmissions and hence the network capacity. Although increasing spatial reuse can enhance the overall network performance, it raises the total interference level on the spectrum, therefore increasing the chance of collisions and adversely impacting the throughput. In this thesis, we optimize the network capacity of IEEE 802.11 based wireless networks through balancing the tradeoff between the level of spatial reuse and collisions in the network. We propose a dynamic spatiotemporal algorithm using the joint control of carrier sensing threshold and contention window size (both are tunable parameters) in order to enhance the spatial reuse and optimize the overall network throughput. In order to efficiently decide which protocol parameter(s) to tune, we deem it necessary to first faithfully distinguish among the causes of frame loss. We propose an effective loss differentiation mechanism to separate packet losses due to collisions from those due to interference(s) and establish accordingly appropriate reactive methods. Simulation results have demonstrated the significant throughput gains that can be achieved by the proposed scheme. We also present decentralized, localized heuristics in order to balance the tradeoff between exposed terminals and hidden terminals, through searching for the optimal transmission power and rate assignment (two other tunable parameters). Our simulation results indicate a remarkable performance for our proposed heuristics.