Data symbol estimation

Abstract

Abstract A method of equalizing a digitised data signal transmitted as a sequence of data symbols. The received signal is generally distorted by transmission through a dispersive communications channel and the detected symbol (D) may differ from the expected symbol (E). The received symbols can exist in a plurality of different states depending on the number of symbols in the data alphabet and the extent of intersymbol interference. An estimate of the data symbol is made based on the maximum likelihood of a particular state transition as represented by the so-called branch metric. In the present method a pseudo-Euclidean branch metric (b) is used which is calculated using exclusively additive arithmetic according to the formula b = |d r - e r | + |d i - e i | where d r and d i are the real and imaginary components respectively of the detected symbol (D) and e r and e i are the real and imaginary components respectively of the expected symbol (E). By avoiding the need for a multiplier the power consumption is reduced. The implementation complexity is also reduced thus minimising the area of semiconductor needed for realization in IC form. A fully multiplier-less equalizer is also disclosed.

Inventors

• 👤 Samuel Ng

Bibliographic Details