Multi-objective evolving neural network supporting SDR modulations management

This paper proposes a distributed Neural/Genetic algorithm able to compute both the more suitable positioning and transmission modulation schemes for fixed/mobile wireless nodes equipped with software defined radio abilities. Devices considered in this work are able to move towards new positions by applying the concept of controlled mobility. The selection of the more suitable modulation scheme is realized through the SDR (Software Defined Radio) paradigm. The synergistic combination of controlled mobility and SDR in a totally distributed way, allows to obtain a high degree of self-configurability; moreover, the extreme adaptability to the network conditions and application level constraints in terms of coverage and guaranteed connectivity, make the proposed approach well suited for quite different communication scenarios such as classical monitoring or disaster recovery. The obtained results, validated throughout an intensive simulation campaign, show how the controlled mobility paradigm applied to the wireless devices and the intrinsic re-configuring SDR capabilities, increase the performance of the network both in terms of coverage and connectivity respect to other algorithms.