Abstract:
This thesis focuses on the application of a parametric approach to the problem of multicomponent polynomial phase signal parameters estimation.
We show how the state-space modellisation and the application of the Kalman filter offer many advantages when combined with the use of multisensor array, exploiting the additional spatial information.
Most conventional approaches are limited to processing narrowband data using narrowband assumption to model the problem.
For tackling the wideband situation, the majority of the proposed methods require transforming the received signals from the time domain to the frequency domain using the Fourier transform as pre-processing step.
Very few authors have dealt with the wideband situation in the time domain.
In this thesis, we recall the definitions of narrowband and wideband signals and their effect on the problem modellisation where we consider a real-valued modellisation, as opposite to the most often used complex-valued modellisation.
The drivers behind this choice are: Dealing with narrowband and wideband signals in the time domain using the same algorithm, and the reduction of the computational cost, as opposite to the complex-valued modellisation, optimising by this the memory use.
The proposed algorithm can be applied in the monocomponent and multicomponent cases with good performance in the case where there are more sources than array sensors.
In addition to the estimation of the signal phase parameters, the resulting algorithm allows the estimation of the sources directions of arrivals and the estimation of the order of the polynomial phase (when unknown).
The proposed algorithm is compared to competitive methods and proven to perform as good or better than these latter with reduced computational cost.
