Patent Application
20070036249
The present invention is generally directed to an FM receiver and, more specifically, to a technique for reducing multipath distortion in an FM receiver having antenna diversity.
As is well known, multipath interference is caused when two or more signal rays of an original transmitted signal converge upon a receiving antenna of a receiver at significantly different times. This misalignment or superposition of signals, which are replicas of the original signal, may cause distortion in audio recovered from the signals. Distortion caused by the multipath interference may be attributable to long delay (e.g., greater than five microseconds between signals) multipath interference or short delay (e.g., less than five microseconds between signals) multipath interference.
In a typical urban environment, RF signals experience changes in amplitude and phase due to short delay multipath. This amplitude and phase shift may result in broadband signal fades of up to 40 dB as the receiver and its associated motor vehicle change locations (see
Generally, long and short delay multipath coexists and creates frequency selective fading and broadband fading, simultaneously (see
For example, as is shown in
In various receiver systems, antenna diversity has been implemented in conjunction with an FM receiver to reduce degraded reception performance caused by multipath interference. Typically, antenna diversity has been accomplished through the use of at least two uncorrelated antennas. Prior art antenna diversity reception for mobile communication systems has been achieved by implementation of a number of different systems. For example, antenna diversity has been accomplished with equal gain combiner (EGC) systems, maximal ratio combiner (MRC) systems and antenna diversity systems, such as the adaptive reception system (ARS) disclosed in U.S. Pat. No. 5,517,686, the disclosure of which is hereby incorporated herein by reference in its entirety.
In general, EGC systems and MRC systems outperform switched antenna diversity systems. However, EGC and MRC systems tend to be more expensive to implement, as they require multiple analog front-ends. Typically, EGC and MRC systems attempt to optimize certain characteristics of a received signal in order to minimize the number of fades or maximize the desired signal.
In a switched antenna diversity system, only one antenna is utilized for reception at any instant in time and, thus, the non-selected antennas do not contribute to the demodulated signal. In contrast, EGC and MRC systems utilize signals from all antennas through a variety of combining techniques.
Constant modulus algorithm (CMA) systems have been implemented extensively in digital broadcasting. In general, combining signals from multiple antennas advantageously provides antenna directionality, which allows interferers (i.e., multipath rays) to be suppressed by creating a null in the antenna pattern in the direction of the multipath interferer. In such systems, the method for determining the combining weights varies, depending upon the application. In the case of an FM receiver implemented in a mobile application, the combining of the signals is particularly challenging, as an implemented algorithm must adapt as the vehicle moves. In a usual case, adaptation is blind, i.e., there is no cooperation between the transmitter and receiver, and the received signal is an analog signal. As such, signal imperfections are difficult to mask.
However, FM modulated signals initially have a constant modulus (amplitude), which provides prior knowledge about the transmitted signal (see
In general, for narrowband modulation, signal combining may be accomplished by adjusting the amplitudes and phases of the individual antenna signals, prior to summation, to optimize appropriate receiver metrics. The gain and phase applied to each antenna signal is equivalent to complex weighting in the complex baseband model and are referred to as antenna or combining weights. The weighted summation of the individual antenna signals results in an overall reception pattern. The combining method determines the weights which optimize certain receiver metrics. In practice, an iterative solution is generally used to solve for the optimal weights and the system is said to ‘adapt’.
The MRC method attempts to maximize some measure of signal-to-noise ratio (SNR) (often called CNR). The CMA method attempts to minimize variation of the received signal amplitude. The CMA method is known to steer the antenna pattern so as to place nulls in the direction of the interfering signals (multipath reflections in this case). The combined antenna pattern varies with the frequency of the received signal so for an FM modulated signal the weights must be adapted rapidly such that the multipath reflection is nulled as it moves in frequency due to the modulation. Because the signal modulation is normally not present in the amplitude of an FM signal, the CMA adaptation bandwidth can extend into the modulation bandwidth such that the adaptation can be fast enough to maintain the necessary null direction as the frequency of the received signal changes due to the modulation.
In general, the CMA method performs well for FM multipath, especially in strong signal conditions where the received distortions are dominated by multipath interference, as opposed to noise. However, as the received signal becomes weak, the amplitude becomes corrupted by noise and the constant modulus assumption of the received signal begins to fail and the CMA performance correspondingly degrades. In addition, the CMA can ‘lock’ to spurs in the received signal that become unmasked as the desired signal becomes weak. In contrast, the MRC adaptation is influenced by the signal modulation and, thus, the adaptation bandwidth is limited. However, the MRC assumptions remain valid even for weak received signals. Therefore, while the CMA tends to outperform the MRC for strong signals, the CMA may have difficulty with weak signals and, in this case, the MRC tends to yield better and more consistent performance.
With reference to
The CMA is a relatively efficient algorithm that allows for adaptation, based on amplitude variation of known constant amplitude signals. In a two antenna case with a single interferer, it is usually possible to position a null in the direction of the multipath interferer (i.e., in the long delay multipath). For an incident plane wave, only the phase of the received signals differ between the two antennas and, thus, combining weights can be found such that the signals cancel in the combined output. While this solution does, in fact, satisfy the constant modulus constraint, it does not guarantee that a system implementing a CMA will adapt to the solution and, in general, the solution may not be unique. For example, a null can be placed in the direction of the direct path and still satisfy the CMA constraint. While this may be acceptable, as noted above, the assumption of constant modulus degrades as the received signal becomes weak, as noise violates the constant modulus property. Furthermore, the wide adaptation bandwidth, which makes the CMA effective in long delay multipath, results in noisy adaptation for weak signals.
What is needed is an improved technique for reducing multipath distortion in a mobile FM receiver having multiple antennas.
According to one embodiment of the present invention, a technique for reducing multipath distortion in a mobile FM receiver, with a plurality of antennas, includes a number of steps. Initially, a first antenna signal and a second antenna signal are received via a first antenna and a second antenna, respectively. Next, the first antenna signal is weighted with a variable first weight and a second antenna signal is weighted with a variable second weight. The weighted first antenna signal is then combined with the weighted second antenna signal to provide a combined received signal having a combined signal level. The variable first and second weights are determined by an adaptation strategy that is substantially controlled by a first algorithm, when the combined signal level is greater than a predetermined signal level. The variable first and second weights are determined by an adaptation strategy substantially controlled by a second algorithm, when the combined signal level is less than the predetermined signal level.
According to another embodiment of the present invention, the transition between the first and second algorithms is achieved by modifying an adaptation step size, when the variable first and second weights are updated. According to one aspect of the present invention, the first algorithm is a constant modulus algorithm (CMA). According to another aspect of the present invention, the first algorithm optimizes reception, when the first and second antenna signals experience multipath interference, where one of the first and second antenna signals is delayed, with respect to the other of the first and second antenna signals, by about 5 microseconds or more.
According to a different aspect of the present invention, the second algorithm is a maximum ratio combiner (MRC) algorithm. According to another embodiment of the present invention, the second algorithm optimizes a signal-to-noise (SNR) ratio of the first and second antenna signals in a fading multipath environment during weak signal conditions, when the signal is less than the predetermined signal level. According to yet another aspect, the second algorithm may be one of an MRC algorithm, an adaptive reception system (ARS) algorithm and an equal gain combiner (EGC) algorithm.
These and other features, advantages and objects of the present invention will be further understood and appreciated by those skilled in the art by reference to the following specification, claims and appended drawings.
The present invention will now be described, by way of example, with reference to the accompanying drawings, in which:
The CMA solution is typically solved iteratively using a gradient descent approach that naturally leads to a least mean square (LMS) like update of the combining weights. The adaptation consists of adding small updates to the combining weights so as to reduce the associated error. The MRC solution can be solved in a similar way such that the weight adaptation consists of updates to the combining weights that minimize the associated error. The updates to the antenna weights are scaled by a parameter called the adaptation step size, usually represented by μ, which controls the rate of convergence and quality of the final solution. By organizing the CMA and MRC solutions to have the identical form, that of LMS-like updates, it is possible to subject the optimal solution to both the CMA and MRC constraints simultaneously by summing both update terms into the weight update. The adaptation step sizes, μCMA and μMRC, then not only control the convergence rate and quality of the solution but the relative emphasis or importance of their corresponding constraint in obtaining the optimal solution. That is, the adaptation effort can be steered between CMA and MRC by adjusting the ratio of μCMA and μMRC.
According to one embodiment of the present invention, a weak signal strategy is employed for robust operation of an FM receiver, with multiple analog front-ends, over a full dynamic range of a received signal. With reference to
As is discussed above, while the constant modulus assumption becomes poor for weak signals, the maximum ratio combiner (MRC) algorithm assumptions remain valid for all signal levels. Furthermore, while the constant modulus (CM) algorithm is generally superior to the MRC at high-received signal-to-noise ratios (SNRs), the MRC algorithm outperforms the CM algorithm for weak noisy signals. According to one aspect of the present invention, an adaptation strategy is implemented that is substantially controlled by the CM algorithm for strong signals and substantially controlled by the MRC algorithm for weak signals. The transition between the two adaptation strategies is accomplished by controlling adaptation step size μ in the weight updates equations, according to the received signal strength.
For strong signals, μCMA>>μMRC, the adaptation is substantially controlled by the CMA. For weak signals, μMRC>μCMA, the adaptation is substantially controlled by the MRC algorithm (see
With reference to the output graphs 1100 and 1200 of
With reference again to the block diagram 900 of
Yk=
and the update equations for the weights are given by:
W—1k=(1−μMRC)W—1k−1+μMRC·X—1k·
W—2k=(1−μMRC)W—2k−1+μMRC·X—2k·
W—Nk=(1−μMRC)W—Nk−1+μMRC·X—Nk·
Accordingly, an FM receiver with multiple analog front-ends has been described herein that exhibits improved performance in the presence of multipath interference.
The above description is considered that of the preferred embodiments only. Modifications of the invention will occur to those skilled in the art and to those who make or use the invention. Therefore, it is understood that the embodiments shown in the drawings and described above are merely for illustrative purposes and not intended to limit the scope of the invention, which is defined by the following claims as interpreted according to the principles of patent law, including the doctrine of equivalents.
