Position-Based Routing in Dynamic Mesh Networks for Applications with Micro-UAVs

Abstract

Many of the ad hoc routing protocols developed in the recent years are intended for the use in large-scale, static networks but provide poor performance in small, dynamic environments with moving nodes. This thesis proposes a new routing algorithm which employs a special position-aided metric to fill this gap. The presented metric rates the wireless links between nodes in the network by predicting the path loss which the radio signals experience during propagation. The path loss is estimated using an empirical log-distance propagation model based on the Euclidean distance between two nodes. The model is maintained live and refitted with new measurements on-the-fly. The resulting link ratings are periodically distributed in the network and used by each node to build and maintain a graph of the network. The graph is used by the individual nodes to determine the most stable paths in the network and to forward messages accordingly. The algorithm is successfully evaluated through multiple outdoor experiments which represent likely scenarios in a dynamic network. Compared to state-of-the-art protocols, it delivers considerably better results in dynamic environments which include moving nodes.