Apparatuses and methodologies for decision feedback equalization using particle swarm optimization

Patent metadata

Field	Value
Publication number	US-9614699-B2
Application number	US-201615066611-A
Country	US
Kind code	B2
Filing date	Mar 10, 2016
Priority date	Aug 12, 2015
Publication date	Apr 4, 2017
Grant date	Apr 4, 2017

Abstract

Methods and apparatuses are provided for channel equalization in a communication system. The method includes initializing, using processing circuitry, filter coefficients of an adaptive decision feedback equalizer randomly in a predetermined search space. Further, the method includes updating, using the processing circuitry, the filter coefficients. The filter coefficients are updated using a least mean square recursion when the filter coefficients are stagnant. The filter coefficients are updated using a particle swarm optimization procedure when the filter coefficients are not stagnant. Further, the updating step is repeated until a predetermined stopping criteria is met. Further, the method includes, filtering, using the processing circuitry, a received signal using the filter coefficients.

First claim

The invention claimed is:

1. A method for channel equalization in a communication system, the method comprising:
initializing, using processing circuitry, filter coefficients of an adaptive decision feedback equalizer randomly in a predetermined search space;
updating, using the processing circuitry, the filter coefficients, wherein in response to determining that the filter coefficients are stagnant, the filter coefficients are updated using a least mean square recursion, and wherein in response to determining that the filter coefficients are not stagnant, the filter coefficients are updated using a particle swarm optimization procedure;
repeating the updating step until a predetermined stopping criteria is met; and
filtering, using the processing circuitry, a received signal based on the filter coefficients,
wherein the updating the filter coefficients using the least mean square recursion includes applying
p i,k+1 =p i,k +μA k T e i ( k )

where μ is a step size, e is an error value,

A
k

=

[

Y
k
H

D
k
H

]

,

 where Y is an input matrix, D is a decision matrix, p i,k is one of the filter coefficients, i is an index, and k is a time instant.

2. The method of claim 1 , wherein the determining whether the filter coefficients are stagnant includes comparing current values of the filter coefficients with previously determined values of the filter coefficients.

3. The method of claim 1 , further comprising: updating the filter coefficients as a function of a velocity update equation, wherein the velocity update equation is a function of an inertia weight, a first acceleration rate, and a second acceleration rate.

4. The method of claim 3 , wherein the velocity update equation includes a constriction factor.

5. The method of claim 1 , wherein the communication system is a multiple input-multiple output communication system.

6. The method of claim 1 , wherein the filter coefficients are complex numbers.

7. The method of claim 1 , wherein the particle swarm optimization procedure includes optimizing an objective function, the objective function being minimized in order to optimize the filter coefficients.

8. An apparatus for channel equalization in a communication system, the apparatus comprising:
processing circuitry configured to
initialize filter coefficients of an adaptive decision feedback equalizer randomly in a predetermined search space,
update the filter coefficients, wherein in response to determining that the filter coefficients are stagnant, the filter coefficients are updated using a least mean square recursion, and wherein in response to determining that the filter coefficients are not stagnant, the filter coefficients are updated using a particle swarm optimization procedure,
repeat the updating step until a predetermined stopping criteria is met, and
filter a received signal based on the filter coefficients,

wherein the updating of the filter coefficients using the least mean square recursion includes applying
p i,k+1 =p i,k +μA k T e i ( k )

where μ is a step size, e is an error value,

A
k

=

[

Y
k
H

D
k
H

]

,

 where Y is an input matrix, D is a decision matrix, p i,k is one of the filter coefficients, i is an index, and k is a time instant.

9. The apparatus of claim 8 , wherein the processing circuitry determines whether the filter coefficients are stagnant by comparing current values of the filter coefficients with previously determined values of the filter coefficients.

10. The apparatus of claim 8 , wherein the processing circuitry is further configured to: update the filter coefficients as a function of a velocity update equation, wherein the velocity update equation is a function of an inertia weight, a first acceleration rate, and a second acceleration rate.

11. The apparatus of claim 10 , wherein the velocity update equation includes a constriction factor.

12. The apparatus of claim 8 , wherein the communication system is a multiple input-multiple output communication system.

13. The apparatus of claim 8 , wherein the filter coefficients are complex numbers.

14. The apparatus of claim 8 , wherein the particle swarm optimization procedure includes optimizing an objective function, the objective function being minimized in order to optimize the filter coefficients.

15. A non-transitory computer readable medium storing computer-readable instructions therein which when executed by a computer cause the computer to perform a method for channel equalization in a communication system, the method comprising:
initializing filter coefficients of an adaptive decision feedback equalizer randomly in a predetermined search space;
updating the filter coefficients, wherein in response to determining that the filter coefficients are stagnant, the filter coefficients are updated using a least mean square recursion, and wherein in response to determining that the filter coefficients are not stagnant, the filter coefficients are updated using a particle swarm optimization procedure;
repeating the updating step until a predetermined stopping criteria is met; and
filtering a received signal based on the filter coefficients,
wherein the updating the filter coefficients using the least mean square recursion includes applying
p i,k+1 =p i,k +μA k T e i ( k )

where μ is a step size, e is an error value,

A
k

=

[

Y
k
H

D
k
H

Assignees

• Univ King Fahd Pet & Minerals

Inventors

• Iqbal Naveed
• Zerguine Azzedine

Classifications

• G06F7/501

  Half or full adders, i.e. basic adder cells for one denomination · CPC title

• H04L2025/03713

  Subspace algorithms · CPC title

• G06F17/16

  Matrix or vector computation {, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization (matrix transposition G06F7/78)} · CPC title

• H04L25/03108

  Theory of recursive equalisers, other than Kalman · CPC title

• H03H17/0277

  comprising recursive filters · CPC title