IS-95, an interim standard proposed for future digital personal communications
systems, uses two levels of encoding of digital data for error control and compatibility
with code-division multiple access (CDMA) transmission. The data is first
convolutionally encoded and the resulting symbols are interleaved and then groups are
encoded as orthogonal Walsh sequences. Decoding these two separate encodings is
traditionally done in separate sequential steps. By combining the decoding and applying
feedback of the final decision of the second level of decoding to the first level decoder
it is possible to reduce the error rate of the decoder.
Each Walsh sequence encodes six non-adjacent symbols of the convolutional code.
The receiver computes an estimate of the probability that each of the sixty-four possible
Walsh sequences has been sent, and uses this estimate as an estimate for each of the
convolution symbols which specified the Walsh sequence. Since the convolution symbols
are non-adjacent, it is likely that the actual value of some of the earlier symbols will have
been determined by the final decoder before later symbols specifying the same Walsh
sequence are used by the convolution decoder. The knowledge of the values of these
symbols can be used to adjust the probability estimates for that Walsh sequence,
improving the likelihood that future convolutional symbols will be correctly decoded.
Specific metrics for estimating probabilities that each convolutional symbol was
sent were tested with and without the proposed feedback, and error rates were estimated
based on extensive computer simulations. It was found that applying feedback does
improve error rates. Analytical methods were also applied to help explain the effects. / Graduation date: 1996
| Identifer | oai:union.ndltd.org:ORGSU/oai:ir.library.oregonstate.edu:1957/34579 |
| Date | 29 February 1996 |
| Creators | Rabinowitz, David |
| Contributors | Magana, Mario E. |
| Source Sets | Oregon State University |
| Language | en_US |
| Detected Language | English |
| Type | Thesis/Dissertation |
Page generated in 0.0019 seconds