One of the primary application scenarios for mobile wireless multi-hop networks are disaster areas. However, these pose specific challenges for routing, such as mobility and highly unpredictable links. The main applications for disaster area networks - group-based voice communication and group-oriented, map-based tracking - can be realized using multicast groups. Thus, we decided to implement ODMRP for disaster area deployments. In several disaster area maneuver on-site deployments, we identified the need for extensions, making the protocol more suitable for real-world deployments. In this paper, we propose three extensions to ODMRP: (1) link quality based routing, (2) prioritization of control messages, and (3) overhead reduction mechanisms. In simulations as well as in real-world measurements, we show the benefit of the extensions proposed.

Wireless multi-hop networks meet the requirements of disaster area scenarios by their definition. Recently, different mesh and Wireless Sensor Network (WSN) testbeds were deployed. However, these deployments do not meet the specific characteristics of disaster area scenarios. Developing algorithms and protocols for public safety scenarios and deploying public safety networks is a huge challenge.

Public safety organizations need robust communication networks to transmit different kind of sensor information. These networks must be reliable even when all infrastructure has been destroyed. Wireless multi-hop networks (such as Mobile Ad-Hoc Networks (MANETs), Wireless Sensor Networks (WSNs), and Wireless Mesh Networks (WMNs)) are supposed to meet the requirements of (1) spontaneous deployment, (2) being independent of any kind of existing infrastructure, and (3) robustness in the sense of self-organization and self-healing by their very definition. These networks have been a topic in research for more than a decade now. Recently, real-world tests and deployments provide valuable insights concerning challenges and future research directions. There are different mesh and WSN testbeds (e.g., [4, 9, 10]) enabling the research community to run tests in static real-world networks. However, concerning public safety requirements, there are significant differences: (1) No spontaneous deployment, (2) no or at least no mobility typical for public safety, (3) no typical applications and traffic for public safety scenarios. Due to these characteristics, developing algorithms and protocols for public safety scenarios and deploying public safety networks is a huge challenge.