Neuro-fuzzy based joint relay-selection and resource-allocation for cooperative networks

Abstract

This paper focuses on a joint relay-selection and resource-allocation algorithm for an Amplify-and-Forward (AF) cooperative network. In a multiuser scenario, joint relay selection and power allocation is a combinational problem for heterogeneous, i.e., real time (RT) and non-real time (NRT), users. In single relay AF (S-AF) scheme, a source-destination pair selects best relay. Thus only two channels are needed (i.e., one for source-destination direct link and other one for the source-relay-destination indirect link) between a source-destination pair. We propose a NeuroFuzzy (NF) based optimal relay selection algorithm for selecting best relay based on link's signal-to-noise ratio (SNR), link's delay and degree of mobility between a source-destination pair. The available radio resources are then allocated sub-optimally to the RT and NRT users on priority basis. The priority parameter depends on Quality-of-service (QoS) requirement of the RT and NRT users. We deduce a close form expression of the moment-generating-function (MGF) for independent and non identical Rayleigh fading channels. Performance evaluations reveal that the proposed joint scheme has lower complexity and better outage behavior as compared to the conventional schemes.