# US 2017/0257234 A1 · Fractionally Spaced Adaptive Equalizer With Non-Integer Sampling

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2017/0257234 A1
- Document ID: US20170257234A1
- Status: Published application
- Kind: A1
- Application number: 15/228,592
- Filed: 2016-08-04
- Published: 2017-09-07
- Earliest claimed: 2016-03-07
- Assignees: Huawei Technologies Co., Ltd.
- Inventors: Syed Faisal Ali Shah; Chuandong Li; Zhuhong Zhang
- Web page: https://patentgrep.com/patent/US20170257234A1

## CPC

- H04L 25/03063: {using fractionally spaced delay lines or combinations of fractionally and integrally spaced taps}
- H04L 25/03949: {equalizer selection or adaptation based on feedback (multiple signaling inclusive of a precoding command for adapting the transmitter H04L1/0031; feedback for transmit diversity systems H04B7/0619; selection of codebook or precoding matrix for MIMO diversity systems H04B7/0456)}
- H04L 25/03885: {adaptive}
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

An apparatus for performing fractionally spaced adaptive equalization with non-integer sub-symbol sampling has an adaptive equalizer that receives a continuous stream of input data having a non-integer, fractional delay between consecutive samples at a non-integer, sub-symbol rate and outputs a stream of equalized data based on tap weights of taps of the adaptive equalizer that are spaced at an interval corresponding to the non-integer, sub-symbol rate. The tap weights are updated independently of the fractional delay between consecutive samples of the input data using an error signal. An equalizer output alignment component downstream of the adaptive equalizer aligns the stream of equalized data with a corresponding transmitted symbol.

## Claims

### Claim 1

An apparatus for performing fractionally spaced adaptive equalization based on non-integer sub-symbol sampling, the apparatus comprising:
an adaptive equalizer configured to receive a continuous stream of input data having a non-integer, fractional delay between consecutive samples at a non-integer, sub-symbol rate and to output a stream of equalized data based on tap weights of the adaptive equalizer, wherein the adaptive equalizer comprises a plurality of taps and a tap weight update component for updating the tap weights using an error signal; and
an equalizer output alignment component disposed downstream of the adaptive equalizer and comprising a fractional delay filter (FDF) for aligning the stream of equalized data with a corresponding transmitted symbol,
wherein the taps are spaced at an interval corresponding to the non-integer, sub-symbol rate, and
wherein the tap weights are updated independently of the fractional delay between consecutive samples of the input data.

### Claim 2 (depends on 1)

The apparatus of claim 1, wherein the equalizer output alignment component further comprises a serial-to-parallel (S/P) converter, wherein the S/P converter is configured to receive the stream of equalized data from the adaptive equalizer and to form a stream of block equalized data comprising a plurality of overlapping blocks.

### Claim 3 (depends on 2)

The apparatus of claim 2, wherein the FDF is configured to receive the stream of block equalized data and to introduce a periodic fractional delay in the stream of block equalized data to time-align each block of the block equalized data with the corresponding transmitted symbol.

### Claim 4 (depends on 3)

The apparatus of claim 3, wherein the FDF processes each block of the block equalized data at a symbol rate and introduces the fractional delay by computing an inner product of the block equalized data with FDF coefficients.

### Claim 5 (depends on 4)

The apparatus of claim 4, further comprising a symbol detector downstream of the equalizer output alignment component, wherein the symbol detector is configured to detect transmitted symbols using the time-aligned equalized data blocks from the equalizer output alignment component and to provide an error signal at the symbol rate.

### Claim 6 (depends on 5)

The apparatus of claim 5, further comprising an error alignment component configured to receive the error signal at the symbol rate from the symbol detector, wherein the error alignment component comprises an FDF to introduce a periodic fractional delay in the error signal to align the error signal with the continuous stream of input data at the sub-symbol rate.

### Claim 7 (depends on 6)

The apparatus of claim 6, wherein the tap weight update component updates one or more tap weights of the adaptive equalizer, and wherein the adaptive equalizer tap weights are updated at the symbol rate using the error signal from the error alignment component and the continuous stream of input data.

### Claim 8 (depends on 5)

The apparatus of claim 5, further comprising an equalizer input alignment component configured to receive the continuous stream of input data at the sub-symbol rate, and wherein the equalizer input alignment component comprises a plurality of FDFs to introduce a plurality of fractional delays in the input data in order to align the input data with the error signal at the symbol rate.

### Claim 9 (depends on 8)

The apparatus of claim 8, wherein the tap weight update component updates one or more tap weights of the adaptive equalizer, and wherein the adaptive equalizer tap weights are updated at the symbol rate using the error signal from the symbol detector and the input data from the equalizer input alignment component.

### Claim 10 (depends on 1)

The apparatus of claim 1, wherein the adaptive equalizer, receiving the continuous stream of input data, is configured to perform short convolutions between the input data and the tap weights to reduce complexity.

### Claim 11

A method of performing fractionally spaced adaptive equalization with non-integer, sub-symbol sampling, the method comprising:
receiving a continuous stream of input data having a non-integer, fractional delay between consecutive samples at a non-integer, sub-symbol rate;
performing an adaptive equalization by an adaptive equalizer of the continuous stream of input data to generate a stream of equalized data at the sub- symbol rate; and
aligning the stream of equalized data with a corresponding transmitted symbol by introducing a fractional delay in the equalized data,
wherein the adaptive equalization of the input data is decoupled from the fractional delay alignment of the input data, and
wherein the adaptive equalization is based on taps that are spaced at an interval corresponding to the non-integer, sub-symbol rate.

### Claim 12 (depends on 11)

The method of claim 11, further comprising performing a serial-to-parallel (S/P) conversion of the stream of equalized data from the adaptive equalizer to form a stream of block equalized data comprising a plurality of overlapping blocks.

### Claim 13 (depends on 12)

The method of claim 12, further comprising introducing a fractional delay in the stream of block equalized data to time-align each block of the block equalized data with the corresponding transmitted symbol.

### Claim 14 (depends on 13)

The method of claim 13, further comprising detecting transmitted symbols using the time aligned equalized data blocks and providing an error signal at a symbol rate.

### Claim 15 (depends on 14)

The method of claim 14, further comprising introducing a fractional delay in the error signal at the symbol rate to provide an error signal at the sub-symbol rate, wherein the error signal at the sub-symbol rate is time-aligned with the continuous stream of input data.

### Claim 16 (depends on 15)

The method of claim 15, further comprising updating, at the symbol rate, a single set of adaptive equalizer tap weights using the time-aligned error signal at the sub-symbol rate and an equalizer block input signal formed by a S/P conversion of the continuous stream of input data.

### Claim 17 (depends on 14)

The method of claim 14, further comprising introducing a fractional delay in the continuous stream of input data to provide a block input data at the sub-symbol rate, wherein the block input data is time-aligned with the error signal at the symbol rate.

### Claim 18 (depends on 17)

The method of claim 17, further comprising updating, at the symbol rate, a single set of adaptive equalizer tap weights using the error signal at the symbol rate and the time-aligned block input data at the sub-symbol rate.

### Claim 19 (depends on 11)

The method of claim 11, wherein the adaptive equalization of the continuous stream of input data comprises short convolutions between the input data and tap weights to reduce complexity.
