next up previous contents
Next: Zero-Forcing Multi-User Detector Up: Fixed Detectors Previous: Conventional Detector (Matched Filter)

Decorrelating Detector

  figure1475
Figure 3.2: Decorrelating DS-CDMA Detector.

The conventional detector does not use any information about the other users in the system and therefore can not combat MAI. The decorrelating detector essentially applies the inverse of the correlation matrix of user spreading codes to the output of the conventional detector, as shown in figure 3.2. This method of decorrelation is only practical at the base-station, where all the information is readily available, and there are no power consumption constraints. The correlation matrix R is defined as
eqnarray419
It is convenient to develop a vector-matrix formulation for the output of the conventional detector. Again assuming no channel dynamics and equal user powers, we can write an expression for our soft decision output vector tex2html_wrap_inline1422
eqnarray428
where
eqnarray421
Now, by applying the inverse of the correlation matrix to the soft decision output vector tex2html_wrap_inline1424, we can completely remove all Multiple Access Interference. The new soft decision of the combined detector is
eqnarray429
As long as the correlation matrix tex2html_wrap_inline1426 is invertible, the soft decision outputs will only be affected by the AWGN. In an asynchronous system, as the uplink in a CDMA system is, tex2html_wrap_inline1426 is assumed to be invertible.

There are some obvious advantages of this detector over the conventional detector. For one, it removes all the multiple access interference. This also implies that the power of each user does not have to be estimated or controlled. One disadvantage to this receiver is noise enhancement caused by tex2html_wrap_inline1430. Also, this detector has no way of eliminating any ISI caused by channel dynamics. Even so, this detector will outperform the conventional detector when the user spreading codes are not orthogonal. The formulation presented does not include asynchronism, but can be extended to do so as shown in [4].

A more general linear detector called the zero-forcing detector eliminates MAI as well as channel ISI, and can be derived from the system model described in chapter 2. Therefore, the decorrelating detector was not implemented in software, but the zero-forcing detector was. Derivation of the zero forcing detector follows in the next section.


next up previous contents
Next: Zero-Forcing Multi-User Detector Up: Fixed Detectors Previous: Conventional Detector (Matched Filter)


Thu Dec 17 13:13:15 EST 1998