Dynamic automatic gain control circuit employing kalman...

Telecommunications – Receiver or analog modulated signal frequency converter – Local control of receiver operation

Reexamination Certificate

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

C455S250100, C330S075000, C327S205000

Reexamination Certificate

active

06766153

ABSTRACT:

FIELD OF THE INVENTION
The present invention relates generally to communication systems and more particularly relates to an automatic gain control circuit incorporating a Kalman filter in the feedback loop.
BACKGROUND OF THE INVENTION
Automatic Gain Control (AGC) circuits are widely used in numerous applications to automatically adjust the level of an input signal to a desired value. Consider a system wherein an amplifier that is part of an AGC circuit functions to generate an output signal by amplifying an analog input signal. The analog output signal is to be converted to its digital equivalent using an A/D converter. Consider also that the input signal is subject to large power variations such as in the case when the signal is derived from a bursty communication medium. In this case, the goal is to properly determine the gain of the amplifier such that the dynamic range of the A/D converter is best utilized.
A block diagram illustrating a prior art automatic gain control (AGC) circuit is shown in FIG.
1
. The AGC circuit, generally referenced
10
, comprises a variable gain amplifier
12
, A/D converter
14
, discriminator
20
, summer
18
and loop filter
16
. The discriminator
20
functions to calculate the logarithm of the output voltage v
out
. This is common practice in AGC circuits to enhance performance by enabling the circuit to work in the logarithmic, i.e. decibel (dB). domain.
The reference quantity R
0
26
is specified by the requirements of the particular AGC implementation and represents the nominal level of the signal at the AGC output. From the point of view of the loop filter, it shifts the gain by a fixed amount. The reference quantity R
0
is subtracted from the output of the discriminator
20
and the difference is input to the loop filter
16
. The loop filter functions to generate the gain g
n
22
.
The purpose of the AGC circuit is to maintain the signal at the input to the A/D converter at an approximately constant level by controlling the gain of the variable gain amplifier placed before the A/D converter.
Typically, a low pass filter is used in the feedback loop in an AGC circuit to reduce the fluctuations in gain (i.e. to generate an average gain). The characteristics of the low pass filter are typically time invariant. Thus, the low pass filter in the AGC feedback loop can be designed to be either wide or narrow, making the AGC either fast or slow. A slow AGC is an AGC that is narrow in terms of frequency. It is better for continuous transmissions since it performs an average of the signal over a longer period of time, is less sensitive to noise and has less gain ripple. A narrow AGC, however, is characterized by poor tracking in that it has more difficulty following quick changes in the level of input signal.
A fast AGC circuit is an AGC circuit that is wide in terms of frequency. It averages the input signal over a shorter period of time, is more sensitive to noise and has a relatively large ripple in its gain.
On the other hand, in burst type communications, it is desirable to keep the gain fixed during the message. It is also desirable to maintain the gain of the signal between packets at a certain level in order to be able to process it. Thus, the AGC circuit is ideally able to quickly adapt to fast changes in the input.
As is the case with most prior art AGC circuits, the bandwidth of the loop gain is much less than that of the signal itself. Variations in loop gain occur much slower than those in the input signal. Note that if the loop gain were equal to the input signal gain, the output of the AGC circuit would be a flat signal.
In many communication systems, it is desirable that the gain of the AGC not change during reception of the information message and during the noise periods as well. This is because the receiver may learn the noise characteristic between packet transmissions. Therefore, adjustments to the gain should occur only during transients between packets and noise periods. This leads to the following contradictory requirements.
The response of the AGC to low to high transitions of the signal level (i.e. the start of a packet) should be very fast. Thus, the AGC feedback loop should be very fast during the noise periods.
The optimum use of the dynamic range of the A/D converter entails maintaining an approximately constant signal level at the output of the AGC circuit. From the above, however, the gain adjustment should only be made during packet
oise period transitions. The precision of the gain setting is thus affected as a result of the constraint of fast adjustment.
Prior art AGC circuits are configured such that their characteristics are fixed, i.e. the circuit operates as a either fast or slow AGC. This, however, may result in unacceptable performance because the requirements described above cannot be satisfied.
A solution to this problem is to detect the beginning of the packet and upon detection to make the AGC loop fast (i.e. wide in frequency) in order to enable quick acquisition. This is followed by a narrowing of the loop gain in order to enable good tracking.
The problem typically encountered in such a system is how to detect the change in input signal level and to optimally perform the transition from wide to narrow loop filtering.
SUMMARY OF THE INVENTION
Accordingly, the present invention provides a novel and useful apparatus for and method of automatic gain control (AGC) using Kalman filtering. The present invention is suitable for use in a wide range of applications and is particularly useful in controlling the gain of the signal received over wired or wireless channels such as power line based channels.
The feedback loop of the AGC circuit of the present invention comprises a Kalman filter augmented by the addition of a non-linear restart mechanism. The circuit is able to transition quickly and make fast adaptations to new levels of the input. It has been found that a Kalman filter is an optimal way in which to estimate the signal. It provides for fast and accurate adaptations to the level of the input signal. The gain of the Kalman filter controls how fast or slow averaging of the signal is performed. A key feature of the present invention is the mechanism used to set the Kalman gain K
n
in an optimal manner. Further, unlike in the prior art, Kalman filtering is used in the present invention in other than the steady state portion of the input signal. Specifically, it is used during acquisition and tracking of the input signal during transient portions of the input.
The input signal is modeled as a stochastic process and the function of the AGC circuit is to try to estimate the level of this signal. The feedback loop comprises a mechanism to detect the level of the signal and use it in generating the gain of the AGC loop.
Three embodiments are presented. The first is an AGC circuit with Kalman filtering which employs a restart mechanism. The restart mechanism enables the AGC circuit to quickly adapt to changes in the signal level of the input. It is used to dynamically modify the gain of the Kalman filter. A restart is triggered when the average of a plurality of previous samples of the AGC output either exceeds an upper threshold or is smaller than a lower threshold. The gain values output by the mechanism can be computed a priori and stored in a table for look up during operation.
It is assumed that the signal power during reception of information packets does not change much as in the case of relatively unchanging channel conditions while the bulk of the gain adjustments occur during transients. The transients are detected by reference to the steady state signal level of the output signal.
A second embodiment comprises an AGC circuit incorporating hysteresis in the feedback loop. The hysteresis circuit is operative to split the loop gain into a feedforward gain and a complementary feedback gain. The hysteresis appears in the feedforward gain only. The feedback gain is used to remove the effects of hysteresis from the AGC output signal before it is fed back to the loop filter, thus making the feedba

LandOfFree

Say what you really think

Search LandOfFree.com for the USA inventors and patents. Rate them and share your experience with other people.

Rating

Dynamic automatic gain control circuit employing kalman... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Dynamic automatic gain control circuit employing kalman..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Dynamic automatic gain control circuit employing kalman... will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFUS-PAI-O-3199242

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.