The present invention relates to signal processing systems and more particularly to improved techniques to reduce interference in radio communication systems.
Recently, receivers have been needed that can operate in different communication networks. Existing and proposed communication networks differ in many ways, including operating on different channel bandwidth specifications and different access technologies for multiple users. The differing processing requirements for modem and protocol functions can be realised with programmable components. These multi-system receivers are usually implemented to cater for the wider bandwidth systems hence there is a need for the front end to handle more number of narrow bandwidth channels falling within the wide channel bandwidth of the wideband system. This could be avoided if programmable bandwidth filters are used to do signal channellization in the front end. But programmable bandwidth filters are hard to realize and using multiple filters catering to different bandwidths will make the system bulk and lossy.
The requirement on ADCs or A/Ds 120 can be reduced by considering the blocking signals as noise overlapping with the required information signal and cancelling the noise out in one of the ways mentioned below. This noise signal can be cancelled out by generating a replica of the blocking signals using prediction techniques and a delayed version of the noise mixed incoming signal and adding in an out-of-phase signal with the incoming signal, as described in U.S. Pat. No. 5,903,819 issued to Romesburg on 11 May 1999. Disadvantageously, this type of scheme requires many components to implement and greater processing power to predict the noise signal at IF frequencies. The validity of the replica sample values depends on how fast the noise estimator works and the accuracy of the estimator.
In some schemes as described in U.S. Pat. No. 3,938,153 issued to Lewis et al. on 10 Feb. 1976 and U.S. Pat. No. 3,938,154 issued to Lewis on 10 Feb. 1976, the interfering blocking signals can be isolated using several filters with different bandwidths and/or downconversion using multiple local oscillators and subtracting from the noise mixed downconverted signal. The drawback of this type of scheme is having a number of local oscillators (LOs), mixers, bandpass filters, and subtractors.
Few implementations for the cancellation of blocking signals have multiple antennas and perform some kind of beam steering to attenuate the interfering signals. Other implementations include demodulating the interfering signal and utilising the demodulation information to neutralise the effects of blocking signals. Examples of the foregoing are described in: U.S. Pat. No. 4,191,926 issued to Pontano et al. on 4 Mar. 1980; U.S. Pat. No. 4,222,051 issued to Kretschmer, Jr. et al. on 9 Sep. 1980; U.S. Pat. No. 4,736,455 issued to Matsue et al. on 5 Apr. 1988; and U.S. Pat. No. 4,384,366 issued to Kaitsuka on 17 May 1983. Any leakage from the transmitter side of a transceiver is cancelled out by extracting a sample of the interfering signal from a coupled signal path of the transmitter and cancelling it out from the incoming received signal after suitable phase detection and adjustments, as described in U.S. Pat. No. 4,660,042 issued to Ekstrom on 21 Apr. 1987.
Alternatively as reported in U.S. Pat. No. 6,169,912 issued to Zuckerman on 2 Jan. 2001, the interfering transmit band signal is extracted from the receive signal itself and used to cancel the interference from the received signal. This type of processing requires some kind of filtering to extract a transmit band signal and is not suitable for suppressing blocking signals that are present in the receive frequency band itself.
The effects of blocking signals on actual information data can be cancelled out in the baseband after data demodulation, as described in U.S. Pat. No. 4,412,341 issued to Geisho et al. on 25 Oct. 1983.
Interference can also be cancelled after elaborately classifying the interference and then mitigating the interference's effects through targeted interference cancellation, as described in U.S. Pat. No. 6,131,013 issued to Bergstrom et al. on 10 Oct. 2000. Though few of these systems are robust, these systems are much easier, or only possible, to implement in the digital domain. This requires high dynamic range ADCs to digitize the signals along with the blocking signals before such processing can be done.
U.S. Pat. No. 3,963,990 issued to Di Fonzo on 15 Jun. 1976 describes cross coupling of signals from two different channels in frequency reuse systems to reduce the interference. However, cross coupling is mainly for interference from channels operating in same frequency but channelized in different polarization angles, and basically not for interference from different frequencies.
U.S. Pat. No. 6,211,671 issued to Shattil on 3 Apr. 2001 interference cancellation schemes for electromagnetic shielding of electromagnetic pickups, other types of electronic equipment, and specific regions of space. This kind of scheme is not suitable for cancelling out blocking signals, as this scheme relies on phase changes in the pickups at different regions in a receiver to effectively cancel out the interference.
Apart from all the cancellation schemes of the incoming blocking signals, this problem may be solved by “near perfect” digital filtering. Several of the foregoing documents have reported improving the dynamic range of the ADCs, so that these signals can be digitized on whole and the required filtering performed to satisfy the blocking tests. However, all of these schemes are disadvantageously complex and consume more power, which are major drawbacks for implementing such schemes in handset kind of applications. Thus, a need clearly exists for an improved technique to reduce interference in receiver architectures of radio communication systems.
In accordance with a first aspect of the invention, there is provided a receiver downconversion architecture for attenuating in an input radiofrequency (RF) signal interfering/blocking signals at offset frequencies from a desired signal. The receiver architecture comprises a delay element having a delay that is dependent on an offset frequency of an interfering signal, and an adder for summing the delayed and instantaneous versions of the input signal.
In accordance with a second aspect of the invention, there is provided a method for, in a receiver downconversion architecture, attenuating in an input radiofrequency (RF) signal interfering/blocking signals at offset frequencies from a desired signal. The method comprising the steps of delaying the input signal dependent on an offset frequency of an interfering signal, and adding delayed and instantaneous versions of the input signal to cancel the interfering/blocking signals.
A small number of embodiments are described hereinafter with reference to the drawings, in which:
A multi-mode receiver/downconverter architecture for use with narrow channel bandwidth and wide channel bandwidth system signals is described. In this architecture, interfering signals for a selected narrowband channel are attenuated using a technique that reduces the dynamic range of the signal for further processing. The technique can be used with receiver architectures, such as direct-conversion, low IF, super heterodyne, and the like. In this technique, the downconverted signal is split into two paths. One signal path is delayed and subtracted from the signal from the other path. By controlling the delay value, the interference signals at a given offset are attenuated. Based on the chosen architecture, the desired signal is placed so that the desired signal undergoes minimum distortion.
The embodiments of the invention attenuate the interfering signals for narrowband systems, which otherwise pass through a wider bandwidth baseband/IF filter catering for the wider bandwidth signals sufficiently enough to reduce the dynamic range requirements of ADCs in the receive chain.
The narrowband signal is quadrature downconverted using an image reject mixer such that desired signal is positioned at a frequency ΔF and it falls at the upper edge of the much wider baseband filter 310 as shown in
A simple way to calculate the amount of delay to be introduced for addition or subtraction of the signals is described hereinafter. Assuming that the information signal is a phase-shift key (PSK) signal S(t) and the interfering/blocking signal is also a PSK signal X(t), the signals can be written
S(t)=A cos [ωct+φc(t)],
X(t)=B cos [ωit+φi(t)],
where: φc(t) and φi(t) are the instantaneous phases;
ωc=2πfc; fc is the required channel carrier frequency;
ωi=2πfi; fi is the interfering channel carrier frequency;
fi=fc+Δf; and Δf is the offset of the interfering signal from the desired signal frequency.
C(t)=S(t)+X(t), and
C
d(t)=C(t−Td).
Assuming that the delay Td is small enough, the phase φc(t−Td) and φi(t−Td) can be approximated to φc(t) and φi(t).
In Cs(t)=C(t)−Cd(t), the information signal adds up and the interfering signal is cancelled out, if:
T
d=1/(2*Δf); and
f
c
=a*Δf; a=1, 3, 5, 7, . . . (‘a’ is an odd integer).
This relationship is valid for direct downconversion, low IF downconversion and the conventional Super heterodyne architecture using higher IF. Importantly the relationship between the carrier frequency and the offset frequency of the interfering signal requiring maximum cancellation is that the carrier frequency should be an odd multiple of the offset frequency of the interference. So that the required signal undergoes 180° phase shift and the interfering signal is phase shifted by zero or multiples 360° phase shift. When the delayed and feed forward paths are subtracted the 180° phase shifted carrier adds up and the other interfering signal cancels out. The scheme can be modified to use a summer instead of a subtractor, in which case the interfering signal will be phase shifted by 180° and the required signal by zero or multiples of 360°.
From the above analysis, by controlling the delay (equal to 1/(2*Δf)) in the feed forward path, the unwanted blocking signals can be cancelled out or attenuated, and the required signals can be added up. The amount of cancellation depends on the amount of phase shift the fixed delay line imparts to the signals and how far the amplitudes of the delayed and instantaneous signals are matched. The amount of cancellation can be quantitatively calculated based on the amplitude and phase error in the two paths. This is also true for instantaneous frequency components in the signal spectrum. For the required signal not to be distorted, the delay has to be sufficiently less than the inverse of the bandwidth of the data modulated on to the required carrier. The delay can be implemented as a fixed delay element for a particular offset frequency to be cancelled or can be made a programmable delay that can be varied to cancel signals at a particular offset.
In this case, the desired signal is downconverted to the IF frequency and the interfering signals at both sides of the IF frequency are attenuated. Depending on the amount of phase shift the delay introduces to the signals, the signals are cancelled or added up.
In the case of the super-heterodyne downconversion architecture, the splitter, delay and subtraction technique (430, 440) is implemented after the IF Bandpass filter (420), which cuts off the far off blocking signals. Depending on the offset frequency of the interfering signal, the fixed/variable delay value Td is calculated. The IF frequency also has to be fixed in such a way that this frequency satisfies the above mentioned conditions.
In the case of quadrature direct downconversion scheme, there is a problem of the image signal, which overlaps with the required signal after downconversion. If the offset frequency is based on the blocking signal, the image frequency signal may be a blocking signal of the narrowband system and be of much higher magnitude than the required signal level. This image frequency has to be removed before any further processing and can be removed by using a image reject mixer architecture.
The module 670 provides at its output the signal at point F of
The use of image reject mixer architecture may not attenuate the image frequency completely and depends on the 90 degree hybrid used to combine the quadrature down converted signal and the signal path lengths after downconversion. As the proposed technique may be used for narrowband signals, the 90 degree hybrid meeting the requirements in the narrow band of interest is sufficient. The attenuation of the blocking/interfering signals leads to the reduction in the dynamic range requirements of the ADC for digitisation and subsequent processing of the multi-mode signals.
The delay Td can be implemented in many ways, examples of which are listed below.
One method uses a simple length of cable or a transmission line having an electrical length that is adjusted so that the cable gives the required delay as calculated above for the required offset frequency. The length of cable can be numerically estimated based on the velocity of electromagnetic (EM) waves in the material in which cable is realised.
Another method integrates the delay into the A/D conversion process. Once the signal is sampled and held, the signal can be split into two paths. One path can be delayed using switched capacitor circuits and combined with the main path samples before quantization. By this way, the limitations on the sample and hold amplifier remain, but the quantizer sees attenuated levels of the blocking signals. The quantizations levels in the quantizer can be set to maximise the dynamic range with reduced number of bits and suitable gain amplifiers can be used to maximise the use of the dynamic range of the quantizer.
Thus, a multimode receiver/downconverter architecture has been described. While only a small number of embodiments have been described, it will be apparent to those skilled in the art that, in the light of this disclosure, modifications and variations can be made without departing from the scope and spirit of the invention.
Number | Date | Country | |
---|---|---|---|
Parent | 10499285 | Aug 2007 | US |
Child | 12469307 | US |