Turbo equalisation of time varying multipath channel under class-A impulsive noise

The paper explores the problem of turbo equalisation and channel estimation under class-A impulsive noise. The channel is a frequency selective fading channel in which its time varying coefficients are expanded into a finite number of basis sequences and time invariant (TI) expansion parameters. Instead of the application of a maximum likelihood (ML) approach in its standard form, the proposed channel estimator is performed by an iterative approach based on the expectation maximisation (EM) algorithm and the steepest descent algorithm. The proposed estimator reduces the complexity of computations resulting from direct application of the ML approach and provides significant performance gain over the algorithms which are efficient in white Gaussian noise. Also, the proposed estimator is suitable for class-A impulsive noise and utilises the soft information obtained from the soft output Viterbi algorithm (SOVA) which is derived under class-A impulsive noise.