Location-quality-aware policy optimisation for relay selection in mobile networks

In order to perform location-based network optimization, the necessary decision information such as current location needs to be collected from the mobile nodes. The accuracy of the decision information, as e.g. influenced by the localization solution, is one of the influencing factors for resulting performance of such network optimizations. In mobile scenarios, the required information collection and forwarding causes delays that will additionally affect the reliability of the collected information and hence will influence the performance of the relay selection method. This paper analyzes the joint influence of these two factors in the decision process for the example of a mobile location-based relay selection approach using a continuous time Markov chain model. Efficient algorithms are developed based on this model to obtain optimal relay policies under consideration of localization errors. Numerical results show how information update rates, forwarding delays, and location estimation errors affect these optimal policies and allow to conclude on the required accuracy of location-based systems for such mobile relay selection scenarios. A measurement-based indoor scenario with more complex mobility and propagation models illustrates the applicability of the model-based policy optimizations for realistic scenarios.