Previous Project


3D-110 Channels

The 3D-110 channel combines MTR modulation coding with a simplified signal space detector implementation of the fixed delay tree search algorithm. To simplify FDTS, an equalization constraint was imposed on the channel so that the first three samples of the equalized impulse response were {1,1,0} while the trailing terms are allowed to take on whatever values minimize the mean squared error at the detector input. Unfortunately for FDTS with tau=2, the minimum distance with this constraint is the same as for the DFE detector. However, by applying an MTR constraint, the symbol distance increases by 3dB. In terms of the constellation, the MTR constraint removes a single symbol that would correspond to three consecutive transitions when combined with the most recent detector output. Figure 1 illustrates the case when the previous decision was +1, so that symbol 5, which represents {-1,+1,-1} is suppressed.

Fig. 1. 3D symbol constellation assuming ak-3=+1. The observation is yk and y'k-2 has the ISI terms due to previous decisions removed.

With the 110 constraint and the MTR code, only three boundaries are required. Each of the resulting linear classifiers combines two of the three observation variables with an offset term and a slicer. A simple AND-OR logic function extracts the decision from the classifier outputs. Simulation results have shown that the 3D-110 detector can match the performance of FDTS/DF tau=2 from normalized user densities from 2.25 to 3.25 or higher, giving it a decided BER performance advantage over competative schemes such as EPRML.

Fig. 2. Block diagram for the 3D-110 detector.


For more information, see

 

Copyright @ University of Minnesota CDSLab
Last modified
Mar 2007