Building scalable and robust wireless mesh networks

City-wide wireless mesh networks are a promising technology for offering ubiquitous Internet connectivity. These multi-hop wireless access networks consist of fixed and mobile nodes, which help each other relaying packets toward Internet gateways and back. The major challenge with these networks is to make them robust and scalable to their size. In this thesis, we therefore develop a scalable routing protocol for robust routing in wireless mesh networks. The first contribution is the design, implementation, and evaluation of HEAT, a scalable and robust field-based anycast routing protocol for routing packets toward the gateways. The fundamental idea of HEAT is based on an analogy from thermal physics. This analogy consists of routing packets along the steepest gradient of a temperature field that is created by the gateways. Furthermore, this temperature field accounts for the link diversity and redundancy toward the gateways. Using simulations, we compare HEAT with two established routing protocols, one reactive and one proactive, in terms of scalability and robustness. The second contribution is the investigation of protocols for routing packets back into wireless mesh networks. Using simulations, we compare three different protocols where each of these protocols represents a routing protocol family. We show that the proactive field-based routing protocol outperforms all