Deployment and performance enhancement of wireless mesh networks

Wireless mesh networks provide necessary network connectivity in rural and remote locations through multi-hopping. To avoid contention and congestion in the wireless channel and degradation in overall performance, a mesh network must be optimized. This dissertation proposes characterization studies in deployment and performance enhancements of wireless mesh networks. Characterizations in wireless mesh network deployments enable network planners to optimize network equipment usage more efficiently. Studies in performance enhancement techniques also enable researchers to create more efficient algorithms. This dissertation illustrates the deployment and performance enhancements characterizations in three phases. The first phase is a large-scale wireless mesh network deployment in a natural reserve. The QuRiNet deployment provides a testing ground for developing new wireless optimization protocols and acts as a backbone network for remote ecological and network data collection. The second phase is a study of wireless link quality in a network. A well-developed wireless link quality characterization study gives insights into the behavior of a wireless network. The final phase uses the link quality study and the multi-radio multi-channel features of QuRiNet for optimization. Specialized static channel assignment algorithms are evaluated to yield a near-optimal network topology. We compare integer linear programming solutions along with breadth-first search and priority-based selection heuristics to determine which algorithm is suitable for QuRiNet.