Browse Prior Art Database

V.9x Modem Constellation Generation Technique

IP.com Disclosure Number: IPCOM000012554D
Publication Date: 2003-May-14
Document File: 6 page(s) / 207K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method for a constellation generation technique for the software V.9x analog modem. Benefits include a reduced symbol error probability for estimated data rates, and a performance improvement over previous methods.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 54% of the total text.

V.9x Modem Constellation Generation Technique

Disclosed is a method for a constellation generation technique for the software V.9x analog modem. Benefits include a reduced symbol error probability for estimated data rates, and a performance improvement over previous methods.

General Description

The disclosed method is based on the super set of network CODEC levels, obtained during DIL sequence processing, the receiver’s equalizer error, and a given power limit. The disclosed method achieves the smallest symbol error probability for estimated data rate, and is verified to gain 1.5% performance improvement over previous methods.

Constellation generation

To provide efficient computations, the whole constellation generation procedure is implemented as set of independent modules:

  • Data rate calculation
  • optimization
  • Constellation masks generation
  • Constellation tuning

These modules are in different sequencer states to avoid the peaky CPU usage. In cases of poor CPU resources, the optimization or constellation tuning or both methods could be omitted.

The whole constellation generation procedure is shown in Figure 1. The following sections briefly describe each of these modules.

Data rate calculation

The method used here is a search through the receiver table with a given � (refer to Zhang et all “Constellation generation and re-evaluation” US Patent 09/572/860). The� new feature of this search is it’s relative rapidity, which is needed because � optimization procedure calls this method many times.

� optimization

The number of levels in the constellation for each slot denoted below as , � and the number of modulus encoder input data bits as . � should satisfy the condition:

        � � � � � � � � � � �                         � � � � � � � � � � � � � � � � � � � � � � � � � � (2.2.1)

In most cases:

        � � � � � � � � � � �         � � � � � � � � � � (2.2.2),

where .

That means the average distance between adjacent constellation points could be increased by decreasing the number of levels , while maintaining the same data rate. A step-type � increasing is used to get its optimum value. The step size can be� chosen as

        � � � � � � � � � � �

the minimum distance between adjacent G.711 points. Figure 2 describes the iteration procedure.

Constellation tuning

Due to discontinuity of V.90 super constellation the technique described above does not allow to make the M-product extremely closed to . In other words, the condition (2.2.2) usually takes place even after optimization. In this case the constellation tuning based on symbol error probability and time slot power distribution is applied. The main idea of this method is to exclude the closest points in the slot with the highest error probability, while maintaining the average constellation power less than given threshold, thus� reducing symbol error rate and improving throughput.

Denote the symbol error probability for each slot as , array of distances between adjacent points in each...