Adaptive Wavelet Modulation for Cognitive Radio

Wavelet Modulation (WM) has been investigated in the past as multi-rate transmission technique on channels unknown in time or bandwidth. Such a channel conditions are common in Dynamic Spectrum Access (DSA) networks where secondary users aim to opportunistically transmit alongside primaries. In this paper, WM is opportunely designed to exploit channel opportunities avoiding interferences with the primary users of the spectrum. Unequal power allocation is performed according to both channel availability and data sensitivity to channel errors: zero power (i.e., carrier suppression) is assigned to potentially harmful wavelets whereas the power budget is distributed on safe wavelets in order to optimize the minimum square error between transmitted and received data. As to optimization Genetic Algorithms (GA) are used due to their capacity to solve complex problems with reduced convergence time and predefined accuracy when tailored to specific cases.