Pulse or digital communications – Spread spectrum – Direct sequence
Reexamination Certificate
1998-07-16
2001-05-08
Le, Amanda T. (Department: 2734)
Pulse or digital communications
Spread spectrum
Direct sequence
C375S150000, C375S227000
Reexamination Certificate
active
06229842
ABSTRACT:
BACKGROUND
In cellular radio systems RAKE receiver structures are used for handling multipath propagation in direct sequence code division multiple access (DS-CDMA) systems. A RAKE receiver should be able to capture most of the received signal energy by allocating a number of parallel demodulators (commonly referred to in the art as RAKE “fingers”) to the selected strongest components of the received multipath signal. After the corresponding delay compensation, the outputs of all fingers are combined. The allocation and time synchronization of the fingers are performed on the basis of the estimated channel response. The multipath delay search processor (commonly referred to in the art as the “searcher”) estimates the channel delay profile, identifies paths within the delay profile, and tracks the delay variations due to changing propagation conditions.
To facilitate demodulation of data transmitted through a radio system using DS-CDMA, the correct code phase(s) of received replica(s) of the transmitted signal must be known at the receiving side. The correct code phase is usually retrieved by the receiver by correlating the received signal with the same, or at least a part of the same, known spreading sequence which was used by the transmitter. The cross-correlation pattern obtained by that operation is then evaluated with respect to relative delay of maxima found in the pattern.
The cross-correlation pattern calculated by the receiver will consist of different types of unwanted signal energy in addition to the desired superposition of cross-correlation values which correspond to the different path delays. This unwanted signal energy is due to appearance of noise and fading in the transmission channel, as well as non-ideal cross-correlation properties inherent with the used spreading sequences. These circumstances will make the cross-correlation peak detection process difficult, since a peak detector may find false correlation maxima (referred to herein as “false alarms”) or may miss existing cross-correlation maxima (referred to herein as “non-detection”).
The problem of finding and retrieving code phase information by detecting cross-correlation maxima has been investigated. A commonly used method, intended to produce a constant false alarm rate, is referred to herein as the constant false alarm rate (CFAR) detector. The principal of the CFAR detector is to provide a path selection threshold value for use in the path estimation such that values above the path selection threshold in the cross-correlation pattern are to be identified as path candidates. If the values fall below the path selection threshold, then the signals are to be rejected and considered as noise. Depending on the value assigned to a threshold value, a certain probability of false path detection, i.e., the false alarm rate, is obtained. Multiplying a predefined constant threshold factor, by the current, measured noise level creates such a path selection threshold value which can be used in a path selection unit to ideally obtain a known, constant false alarm rate. The constant threshold factor used in this conventional detector may be optimized for a given set of system operating parameters and conditions.
Closely connected to the choice of the threshold factor, and the corresponding probability of false alarm detection, is the probability of not detecting existing cross-correlation maxima, i.e., the non-detection rate. If the path selection threshold is set at a relatively high level, then the number of false alarms decreases, but the number of non-detections increases. Conversely, if the path selection threshold is set at a relatively low level, then the number of false alarms increases, but the number of non-detections decreases. Since minimization of both the non-detection and false alarm probabilities are desirable for overall receiver performance, and because the minimization of these probabilities raise contradictory requirements regarding the setting of the detector path selection threshold, a careful setting of this path selection threshold is important for any system applying this method of path searching.
FIGS. 1A and 1B
provide a conceptual illustration to aid in the understanding of how setting the path selection threshold to minimize both false alarms and non-detections can, at times, create contradictory requirements.
FIG. 1A
illustrates the probability of detecting false paths and non-detection of valid paths where there is a high signal-to-noise ratio (SNR). As can be seen in
FIG. 1A
, when a constant false alarm rate path selection threshold (th
CFAR
) is used for peak detection, although the probability of non-detection is zero the CFAR detection unit has some fixed, and constant, false alarm rate. Further,
FIG. 1A
shows that under these SNR conditions, moving the threshold value to the point illustrated as th
adaptive
would result in no false path detections or non-detections of valid paths.
FIG. 1B
illustrates the probability of detecting false paths and non-detection of valid paths when there is a low SNR. In this Figure, it can be seen that the use of a constant path selection threshold, during periods of low signal-to-noise ratio, results in an increased probability of non-detection, while maintaining the substantially fixed probability of false alarms. Further,
FIG. 1B
shows that under these SNR conditions, moving the threshold to the left would minimize non-detections at the expense of an increased false alarm rate, as illustrated by the adaptive threshold, th
adaptive
, that tradeoff between non-detections and false alarms may be desirable as described below.
The graphs illustrated in
FIGS. 1A and 1B
are purely conceptual and used to point out that Applicants have discovered that the traditional algorithm, which adapts the path selection threshold used in the peak detector by multiplying the mean noise level with a constant threshold factor, does not result in optimum overall receiver performance. Although the constant threshold factor used in determining the path selection threshold may be optimal for a given set of operating parameters and conditions, this constant threshold factor is not optimal for other parameters and conditions. Thus, for good transmission conditions (e.g., high signal-to-noise ratios (SNR's)) the conventional algorithm might detect false correlation peaks which results in a degraded overall performance. For bad transmission conditions (e.g., low SNR's) the conventional algorithm is conservative (i.e., the threshold is too high) and it will reject potential correlation peaks, which may deteriorate the overall receiver performance. Hence, these observations indicate that any chosen constant threshold factor will not optimize the overall performance of the receiver and thus not optimize the capacity of a system.
SUMMARY
Systems and methods for determining valid paths in a spread spectrum radio receiver are described. According to an exemplary embodiment of the present invention, a path selection unit is used to determine valid peaks in a delay power profile (DPP) signal. The present invention accurately estimates the noise level present in DPP signals in order to create an accurate SNR estimate which is mapped using a threshold table or a mapping function to adaptively set a path selection threshold for separating valid peaks from noise in the DPP signal. Using accurate noise level estimates and adaptive path selection threshold values, optimizes the probability of both non-detections and false alarms.
According to an exemplary embodiment of the present invention, the system continuously estimates the channel SNR to set the path selection threshold. The SNR is estimated using an iterative process of determining the noise level present in the signal. The iterative process includes the step of determining a raw estimate of the noise level by removing a predetermined number of peaks and evaluating the residual signal. Then, an improved noise estimate is determined by using the raw noise level estimate to refine the number of peaks that shoul
He Ning
Klang Göran
Schulist Matthias
Burns Doane Swecker & Mathis L.L.P.
Le Amanda T.
Telefonaktiebolaget LM Ericsson (publ)
LandOfFree
Adaptive path selection threshold setting for DS-CDMA receivers 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 path selection threshold setting for DS-CDMA receivers, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Adaptive path selection threshold setting for DS-CDMA receivers will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2456570