Mobility Impact on Centralized Selection of Mobile Relays

This paper considers the impact of node mobility on the selection of the lowest bit error rate (BER) relay path from access point to destination node in a downlink data transmission scenario. The access point uses link signal to noise ratio (SNR) measurements provided by the mobile nodes to select the path either as direct or two-hop via a mobile relay. The impact of delays in the measurement collection procedure in this scheme is analyzed in comparison to an ideal selection and an always direct scheme. An ns-2 and matlab based simulation framework is used to investigate effects of varying scenario parameters on the achieved average BER of data transmissions. The SNR measurement collection is based on periodic hello broadcasts.
Results show that increasing node speed can lead to performance that is worse than always transmitting directly. The hello broadcast rate can mitigate this effect, however, at the cost of an increase in signaling overhead. We find that discarding old measurements can significantly improve the BER performance without increasing the signaling overhead. This result establishes the importance of joint tuning of the hello broadcast rate and the measurement storage time.