Pulse or digital communications – Spread spectrum – Direct sequence
Reexamination Certificate
2001-07-03
2004-04-27
Vo, Don N. (Department: 2631)
Pulse or digital communications
Spread spectrum
Direct sequence
C375S149000, C375S354000
Reexamination Certificate
active
06728303
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to communications. Particularly, the present invention relates to timing discrimination in a communications environment.
2. Description of the Related Art
A conventional Rake receiver operates on received signals with correlators known as “fingers”. A Rake receiver, on the basis of complex channel coefficients of each desired multipath component, coherently combines finger outputs. Complex channel coefficients may be obtained from periodically transmitted training sequences, in a known manner.
In code division multiple access (CDMA) systems, it is known to assign multiple fingers to different signal paths. Each finger tracks its assigned path using a closed loop, time tracking loop (TTL) associated therewith to control its location.
When the paths become sufficiently close to each other, the TTL's of the fingers tracking individual paths sometimes cause the lose fingers to merge. Merging, due to adverse fading conditions, implies, for example, one finger abandons the path it was originally tracking and starts tracking a path already being tracked by another finger.
To avoid fingers from merging, it is known to use a simple early−late timing discriminator, whose output enables the associated TTL to “pull-in” a path that may be up to one chip away from the current position of the finger. This concept is illustrated in FIG.
1
.
FIG. 1
shows a plot of the output of a receive filter versus time. It would be best to sample the waveform at the peak during the on-time sample. However, since this cannot always be accomplished, an early sample is taken approximately half a chip time (Tc/2) before the on-time sample and a late sample is taken approximately half a chip time after the on-time sample.
If the early sample minus the late sample (E−L) is equal to zero, then &Dgr;t
res
=0. If E−L<0, then &Dgr;t
res
<0 showing that the sample was taken too early. If E−L>0, then &Dgr;t
res
>0 showing that the sample was taken too late.
The Early−Late timing discriminator works only as long as the paths are far enough apart such that the curve of
FIG. 1
for the first path does not overlap the same curve for a second path. Due to fading, there may be a time when one of the paths is much weaker than the other one. When this happens the timing discriminators of both fingers will be pulling both fingers towards the strongest path. If this condition lasts long enough the finger that used to be tracking the weak path will start to move in the direction of the strong path. Eventually, fading conditions will move the fingers close enough together until they merge. Once the fingers merge they cannot be separated because their timing discriminator outputs will be identical.
The tracking of a single path by multiple fingers is an undesirable condition for many reasons. Two or more fingers are now being wasted on one path. Also, the energy from the “abandoned” path is lost because the finger that was tracking this path is now merged with another path.
Merged fingers also degrade performance in more subtle ways. For example, the merged fingers not only do not improve the performance of the Rake receiver but also will contribute to the
E
c
N
t
⁢
⁢
or
⁢
⁢
E
c
I
o
estimation used for power control. This results in a mobile station requesting the base station to increase transmit power. If there are any other fingers combining, this disproportionately weights the paths with merged fingers tracking them.
By avoiding Rake finger merging, the performance of the overall system is enhanced as measured by the frame error rate (FER) of the system as a function of the signal to noise ratio. There is a resulting need to prevent Rake finger merging and allow the fingers to freely track their paths in order to greatly reduce the frequency that the processor must interfere with the operation of the fingers.
SUMMARY OF THE INVENTION
The present invention encompasses a process and apparatus for timing discrimination that prevents Rake finger merging. The process for timing discrimination occurs in a timing discriminator that is part of a time tracking loop. The timing discriminator is coupled to an input signal.
The process begins by gathering early, ontime, and late samples of the input signal. Early, Ontime, and Late parameters are then derived in response to the samples.
The timing discriminator output is generated in response to a predetermined relationship between the Early, Ontime, and Late parameters. In the preferred embodiment, this relationship is expressed as
if (&bgr;·Early>Ontime) OR (&bgr;·Late>Ontime)
TD=0
otherwise
TD
=Early−Late
The present invention, therefore, enables Rake fingers to freely track their paths. This improves the receiver performance.
REFERENCES:
patent: 5903308 (1999-05-01), Cooper et al.
patent: 5960040 (1999-09-01), Cai et al.
patent: 5978423 (1999-11-01), Farjh
patent: 6078611 (2000-06-01), La Rosa et al.
Brown Charles D.
Louis Timothy F
Qualcomm Incorporated
Vo Don N.
Wadsworth Philip R.
LandOfFree
Timing discriminator with merge protection does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Timing discriminator with merge protection, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Timing discriminator with merge protection will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3236640