Electrical computers: arithmetic processing and calculating – Electrical digital calculating computer – Particular function performed
Reexamination Certificate
2000-04-26
2004-01-06
Ngo, Chuong Dinh (Department: 2124)
Electrical computers: arithmetic processing and calculating
Electrical digital calculating computer
Particular function performed
Reexamination Certificate
active
06675184
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention:
The present invention relates to an adaptive type signal estimator and, more particularly, to a signal estimator for reducing estimation errors in estimating a transmission signal sequence from a reception signal influenced by noise, fading that changes over time, and the like in an environment in which multipath distortion exists as in mobile communication. A reception signal to be processed is, for example, a signal accompanied by nonlinear distortion such as in narrow-band GMSK (Gaussian filtered Minimum Shift Keying).
2. Description of the Prior Art:
As a conventional device for estimating a transmission signal sequence from a nonlinearly distorted reception signal, an automatic equalizer using a RAM (Random Access Memory) is available.
FIG. 1
is a block diagram showing such an automatic equalizer, which is disclosed in Japanese Unexamined Patent Publication No. 5-14126 (“automatic equalizer”). Binary data of (+1, −1) is transmitted from the transmitting side upon GMSK modulation, and the signal influenced by multipath distortion in a transmission path is received on the receiving side. In general, GMSK modulation is a nonlinear modulation scheme which limits the signal band through a Gaussian filter. For this reason, this scheme produces intersymbol interference containing nonlinear distortion in a modulated signal. As the frequency band of a modulated signal is narrowed, in particular, nonlinear distortion increases. In mobile communication, a signal is influenced by multipath distortion in a transmission path. This multipath distortion leads to intersymbol interference due to linear distortion. On the receiving side, therefore, a signal containing both linear distortion and nonlinear distortion is received. In this case, as a distortion model, intersymbol interference spread across five symbol intervals is used. That is, a reception signal x
n
is influenced by five transmission data sequences (u
n+2
, u
n+1
, u
n
, u
n−1
, u
n−2
). This phenomenon can be mathematically expressed as x
n
=f[u
n+2
, u
n+1
, u
n
, u
−1
, u
n−2
]. A function f is a function containing both linear distortion and nonlinear distortion.
Referring to
FIG. 1
, an estimation signal storage unit
204
generates all signals that can be received in a noiseless condition as sample signals. That is, 32 types of sample signals represented by y
31
=g[+1, +1, +1, +1, +1] are generated from y
0
=g[−1, −1, −1, −1, −1]. These signals are stored in a programmable memory typified by a RAM. The function g is a function that is obtained by estimating the function f on the receiving side.
An error detector
201
obtains 32 types of branch metrics required for arithmetic operation based on a Viterbi algorithm from the 32 types of sample signals (y
0
to y
31
) and the reception signal (x
n
).
A comparator
202
determines a signal by performing arithmetic operation based on the Viterbi algorithm using the 32 types of branch metrics output from the error detector
201
.
In general mobile communication, since users communicate with each other on move, multipath distortion also changes over time. That is, the function f changes with time. Therefore, the sample signals stored in the estimation signal storage unit
204
must be changed in accordance with changes in multipath distortion over time. The sample signals are corrected as follows.
To correct the sample signals stored in the estimation signal storage unit
204
, a corrector
205
and address generator
203
are required. The address generator
203
sequentially stores the data determined by the comparator
202
, and generates an estimated value (w
n+2
, w
n+1
, w
n
, w
n−1
, w
n−2
) of transmission data u
n
. In accordance with the output from the address generator
203
, the estimation signal storage unit
204
outputs y
n
=g[w
n+2
, w
n+1
, w
n
, w
n−1
, w
n−2
] as a sample signal. The corrector
205
then corrects the sample signal y
n
so it approaches the reception signal x
n
. By performing this operation sequentially, the 32 types of sample signals stored in the estimation signal storage unit
204
gradually change. This reduces to the function g always tracking changes in the function f.
As described above, since the distortion contained in a reception signal changes with time, sample signals must always be corrected in accordance with the changes in distortion. As a method for this correction, a method of correcting a sample signal by comparing the sample signal obtained from a determination signal with a reception signal is available. In this method, since only one sample signal can be corrected by using one reception signal, an average of 32 reception signals must be input to correct all the 32 types of sample signals.
A scheme that cannot cope with nonlinear distortion, e.g., a scheme using a transversal equalizer, can correct all sample signals by using one reception signal, and hence can track changes in distortion over time no matter how the changes increase. In this scheme, however, since the tracking speed decreases to 1/32, the tracking performance greatly deteriorates.
In addition, as intersymbol interference is prolonged, the number of sample signals exponentially increases. For this reason, the hardware size increases, and the tracking speed decreases exponentially as well.
SUMMARY OF THE INVENTION
The present invention has been made in consideration of the above situation in the prior art, and has as its object to provide an adaptive type signal estimator which receives a signal accompanied by nonlinear distortion, and can quickly estimate a transmission signal sequence even if the reception signal is influenced by noise, fading that changes over time, and the like.
In order to achieve the above object, according to the main aspect of the present invention, there is provided an adaptive type signal estimator comprising an estimation signal storage unit for outputting a first sample signal to which nonlinear distortion is added, a convolution arithmetic unit for adding linear distortion to the first sample signal and outputting the signal as a second sample signal, a branch metric arithmetic unit for generating a branch metric used for Viterbi algorithm operation from a reception signal and the second sample signal, and outputting the branch metric, a signal determination unit for outputting a determination signal by performing maximum-likelihood sequence estimation using an output from the branch metric arithmetic unit on the basis of a Viterbi algorithm, and a coefficient corrector for receiving the first sample signal, the reception signal, and the determination signal, and correcting the second sample signal output from the convolution arithmetic unit, wherein the coefficient corrector obtains a difference signal on the basis of a replica generated from the reception signal and the first sample signal and a delayed reception signal obtained by delaying the reception signal by a predetermined period of time, obtains a product signal by multiplying the difference signal by a convergence factor, updates an impulse response value by using the product signal, and outputs the updated impulse response value to the convolution arithmetic unit.
According to the present invention, in obtaining sample signals used to generate branch metrics in a Viterbi algorithm, nonlinear distortion caused in a modulator on the transmitting side and linear distortion caused by a multipath effect in a transmission path are taken into consideration separately to perform adaptive control on only linear distortion that changes over time so as to track only changes in linear distortion over time, but no adaptive control is performed on nonlinear distortion that does not change over time.
To obtain a sample signal, a signal that does not change over time is generated by using a memory table i
Dickstein Shapiro Morin & Oshinsky LLP.
Do Chat C.
NEC Corporation
Ngo Chuong Dinh
LandOfFree
Adaptive type signal estimator does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Adaptive type signal estimator, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Adaptive type signal estimator will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3199413