Channel noise is stationary, additive and white with infinite bandwidth.

Fast fading can be minimized by using coding and interleaving, by using robust modulation and by introducing signal redundancy to increase signalling rate.

Frequency selective distortion can be minimized by using adaptive equalization, by using spread spectrum, by using pilot signal and also by using orthogonal FDM modulation.

Only the relevant noise components play an important role in decision making. It does not depend on non relevant noise components.

The inter-leaver is more effective in fast running vehicles.

The known ISI which is introduced deliberately and the adjacent channel interference is removed by cancelling filter and post processing equalizer.

Slow fading can be minimized by using error correcting codes and also by using diversity technique to get additional uncorrelated estimates of a signal.

Fading channel has memory and the received samples are correlated with each other in time.

Decision making needs both priors and likelihoods and Bayes decision rule combines them to achieve minimum probability of error.

The decision feedback filter has a feed forward section which is a linear traversal filter.