Browse Prior Art Database

AN OPTIMAL METHOD FOR CHOOSING ENERGY QUANTIZATION LEVELS FOR TRELLIS SOFT DECISION DECODING OF DATA

IP.com Disclosure Number: IPCOM000007325D
Original Publication Date: 1995-Mar-01
Included in the Prior Art Database: 2002-Mar-14
Document File: 1 page(s) / 71K

Publishing Venue

Motorola

Related People

Kevin Doberstein: AUTHOR

Abstract

Existing data systems employing sofi decision decoding are subject to short RF noise bursts that can adversely affect decoding at "moderate" signal strengths. These bursts, even though they are only one to two symbols times in length, cause the decoder to fail since the burst affects both parame- ters used in the decoding, the symbol value and the corresponding instantaneous symbol energy. The symbol value is not valid, but the since the energy level is raised due to the burst, the decoder to "weighs" the incorrect symbol heavier than the cor- rect symbols causing the decoder to fail.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 58% of the total text.

Page 1 of 1

0 M

MOTOROLA Technical Developments Volume 24 March 1995

AN OPTIMAL METHOD FOR CHOOSING ENERGY QUANTIZATION

LEVELS FOR TRELLIS SOFT DECISION DECODING OF DATA

by Kevin Doberstein

  Existing data systems employing sofi decision decoding are subject to short RF noise bursts that can adversely affect decoding at "moderate" signal strengths. These bursts, even though they are only one to two symbols times in length, cause the decoder to fail since the burst affects both parame- ters used in the decoding, the symbol value and the corresponding instantaneous symbol energy. The symbol value is not valid, but the since the energy level is raised due to the burst, the decoder to "weighs" the incorrect symbol heavier than the cor- rect symbols causing the decoder to fail.

  Current implementations of the energy quanti- zation look at the whole range of energy values and quantize based on the entire range. This leaves the algorithm open to potentially weighing a symbol corrupted by a noise burst under moderate to strong signal strengths more heavily that it should, thus causing the decoder to fail.

  This new quantization method uses the concept that the no particular symbol should be weighed heavier than the next while the signal strength is such that the receiver can recover a signal with 0% BER. This amounts to a non-linear quantization scheme for the energy values. Figure 1 shows two different representations of the energy, one with no scaling and one scaled by 2A 14. The scal...