WO1994000918A1 - Procede et appareil d'annulation d'interferences dans un recepteur de trafic a acces multiple - Google Patents

Procede et appareil d'annulation d'interferences dans un recepteur de trafic a acces multiple Download PDF

Info

Publication number
WO1994000918A1
WO1994000918A1 PCT/US1993/006125 US9306125W WO9400918A1 WO 1994000918 A1 WO1994000918 A1 WO 1994000918A1 US 9306125 W US9306125 W US 9306125W WO 9400918 A1 WO9400918 A1 WO 9400918A1
Authority
WO
WIPO (PCT)
Prior art keywords
signals
signal
interference
amplitude
set forth
Prior art date
Application number
PCT/US1993/006125
Other languages
English (en)
Inventor
George A. Zimmerman
Edward C. POSNER
Original Assignee
Sotel, Phillip, K.
POSNER, Sylvia
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Sotel, Phillip, K., POSNER, Sylvia filed Critical Sotel, Phillip, K.
Priority to AU46533/93A priority Critical patent/AU4653393A/en
Publication of WO1994000918A1 publication Critical patent/WO1994000918A1/fr

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B1/00—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
    • H04B1/06—Receivers
    • H04B1/10—Means associated with receiver for limiting or suppressing noise or interference
    • H04B1/12—Neutralising, balancing, or compensation arrangements
    • H04B1/123—Neutralising, balancing, or compensation arrangements using adaptive balancing or compensation means
    • H04B1/126—Neutralising, balancing, or compensation arrangements using adaptive balancing or compensation means having multiple inputs, e.g. auxiliary antenna for receiving interfering signal

Definitions

  • This invention relates generally to improvements in methods and apparatus for increasing the capacity of multiaccess communication systems, such as cellular telephony, portable telephones, mobile radio communications, broadcast radio and television, wireless communication networks and the like, and, more particularly, to a new and improved system for harnessing interfering signals to increase system capacity for multiple users.
  • multiaccess communication systems such as cellular telephony, portable telephones, mobile radio communications, broadcast radio and television, wireless communication networks and the like
  • CDMA Code Division Multiple Access
  • each signal uses a much larger bandwidth than is required to transmit the desired information, and a coding scheme to choose the desired signal from multiple signals in the same bandwidth.
  • This requires rather complicated receivers which must receive much more bandwidth than is normally required for each user.
  • Some forms of CDMA are also known as spread spectrum multiple access, in which the wide-band codes make interfering signals appear to be white noise to the desired signal receiver.
  • the required bandwidth of the desired signal is greatly expanded.
  • the desired signal's bandwidth is collapsed back to the information-bearing bandwidth, while the interfering signal's bandwidth remains spread.
  • the interference is not cancelled or removed. It still reduces the network capacity by contributing to the noise.
  • the information-bearing structure of the interference is not used in any way to mitigate this undesirable effect.
  • Prior art systems dealing with interference generally tend to cancel interference to receive a single selected signal, rather than harness interfering signals to provide increased capacity in a multiaccess system.
  • Such prior art approaches attempt to mitigate the effect of a generally weaker interference signal on the reception of a generally stronger desired signal.
  • the interference in such prior art system is typically totally discarded. Modulation or information contained in the interference is generally not used in any way in its mitigation.
  • prior art systems require a second channel, a so-called "reference" input of the interference, in order to cancel the interference.
  • the prior art may separately demodulate a desired signal, but does not attempt to improve overall network capacity by jointly demodulating multiple, interfering signals.
  • FM frequency-modulated
  • SNR signal-to-noise ratio
  • the FM capture effect facilitates a simple method of providing multiaccess communications by dividing the service area into geographical regions.
  • the allocation of frequencies for FM transmitters has been traditionally accomplished in this manner.
  • the allowable geographical separation of co-channel and adjacent channel FM transmitters be they high-power broadcast stations or relatively low-power cellular telephones, has been regulated so that propagation laws will mitigate any interference problems.
  • intereference-cancelling receivers in accordance with the present invention, to such cellular channels will decrease the minimum frequency reuse distance, allowing the system to carry more traffic on the same allocated spectrum, i.e., more channels per cell.
  • a change in the cellular system could be implemented without the lengthy and expensive regulatory procedures required for changing the modulation schemes or obtaining additional bandwidth.
  • a commercial broadcast station can be considered a multiuser communications system over a large geographical area that includes multiple stations assigned to the same frequency. Most of these systems face much less severe reception environments than the cellular telephone channel. In more general mobile communications applications, the required level of voice quality can be far below cellular telephone standards, e.g., on a police, aircraft, or taxicab radio.
  • Application of interference cancellers to these channels, in accordance with the present invention, would not only allow several users to speak at once, something not possible on current systems, but would also allow the allocated bandwidth to support more user licenses. Application to these mobile channels would be similar to the cellular telephone channel, since the cellular telephone channel is also a mobile channel.
  • the present invention provides a new and improved multiaccess communication system embodying novel methods and apparatus for extracting a selected signal from a larger family of such signals having some relationship to the selected signal and overlapping predictable electrical characteristics wherein a desired signal to be received is selected from the family of signals, an estimate is made of the magnitude of at least one predetermined electrical parameter in the remaining unselected signals of the family which would normally interfere with the selected signal, and the estimate is used to substantially cancel the unselected signals and facilitate enhanced reception of the desired signal.
  • any signal in the family of signals may be designated as the desired signal to be received to thereby increase system quality, capacity and availability to an increased number of users.
  • what would normally be interference is selectively harnessed to increase the availability of overlapping signals to multiple users.
  • interference cancellation in accordance with the current invention, removes interference, guaranteeing enhanced reception of the desired signal, and is robust to these propagation effects.
  • planning and designing to allow such interference to increase capacity for multiaccess is entirely new and represents a breakthrough for expansion of communications systems without the need for expanded frequency and/or spatial diversity.
  • Network capacity can be increased if multiple signals can overlap in transmission parameters, e.g., frequency or polarization.
  • transmission parameters e.g., frequency or polarization.
  • the prior art often relies on filtering a separable physical transmission parameter, such as frequency or polarization.
  • Inadequate initial separation, e.g., frequency overlap or insufficient cross-polarization isolation result in system degradation.
  • the system of the present invention is freed from such constraints. Signals can overlap in frequency, and need not be separated by polarization.
  • a second "reference" channel input is not needed in the system of the present invention to cancel the interference.
  • a channel which includes only the interference and not the desired signal, is not available with only a single antenna and receiver.
  • the system of the present invention uses the inherent information redundancy present in the modulation format, the modulating signal, or the channel path-induced distortions of almost all practical systems to separate interfering signals and, hence, does not add additional bandwidth to the signal at transmission.
  • the estimation of an interfering waveform is used to cancel interference in the form of undesired signals wherein any or all of the signals may be information bearing signals or constructed signals having known and predictable characteristics, as opposed to "white noise" having totally unpredictable characteristics.
  • Such interference cancellation yields multiple selective access to the pool of signals for extraction of desired signals by multiple users.
  • the aforedescribed invention may be embodied in various analog and digital receiver configurations, utilizing cross-coupled phase locked loop technology, in configurations embodying inphase, inphase and quadrature, feed forward and difference amplitude-tracking electrical networks.
  • the new and improved multiaccess communication system of the present invention is unique in that it harnesses interference to increase system capacity available for plural access by single or multiple users.
  • the improved system provides receivers which are independent of any power hierarchy in that desired signal extraction does not depend on relative position of the desired signal in the power hierarchy.
  • the system can detect the weakest signal in the family of overlapping signals, and the relative power hierarchy need not be stable or preserved.
  • the present invention expands the useful bandwidth in multiaccess systems and satisfies a long existing need in the communications industry for accurate, reliable, easy to use and economical systems for exploiting the frequency spectrum to provide increased system capacity for multiple users.
  • FIG. 1 is an overall system block diagram of a network system incorporating features of the present invention to achieve increased capacity
  • FIG. 2 is a block diagram of a simple receiver subsystem embodying the invention
  • FIG. 3 is a block diagram illustrating an alternative embodiment of a receiver subsystem in accordance with the invention.
  • FIG. 4 is a block diagram of another embodiment of a receiver subsystem in accordance with the invention.
  • FIG. 5 is a conceptual block diagram of an FM Multi-User Receiver embodying features of the invention.
  • FIG. 6 is a block diagram illustrating independent interference estimation for plural interfering signals
  • FIG. 7 is a block diagram illustrating vector sum estimation in a receiver subsystem
  • FIG. 8 is a block diagram illustrating a basic Cross-Coupled Phase Locked Loop receiver configuration
  • FIG. 9 is a block diagram illustrating Estimator-Subtractor Interference Cancellation for a receiver
  • FIG. 10 is a block diagram illustrating an Optimum (MAP, or Maximum A Posteriori) Receiver for FM with interchannel interference;
  • FIG. 11 is a block diagram illustrating a CCPLL (cross-coupled phase locked loop) with Amplitude and Leakage Control;
  • FIG. 12 is a block diagram illustrating a receiver subsystem embodying a Feedforward (Direct Form) Optimum Amplitude Estimator;
  • FIG. 13 is a block diagram illustrating a receiver subsystem embodying a Feedback (Gradient-Descent Form) Optimum Amplitude Estimator
  • FIG. 14 is a block diagram illustrating a receiver subsystem embodying Difference-Amplitude Tracking configuration in accordance with the invention
  • FIG. 15 is a basic block diagram illustrating a digital system embodying a presently preferred embodiment of the invention.
  • FIG. 16 is a block diagram of an analog to digital conversion subsystem suitable for use with the system of Fig. 15;
  • FIG. 17 is a block diagram illustrating an interference cancelling FM demodulator suitable for use with the system of Fig. 15;
  • FIG. 18 is a block diagram for a phase detector suitable for use with the system of Fig. 15;
  • FIG. 19 is a block diagram illustrating the loop gain and filter for the system shown in Fig. 15;
  • FIG. 20 is a block diagram illustrating an amplitude estimator suitable for use with the system of Fig. 15;
  • FIG. 21 is a block diagram illustrating a 1st Order Digital Butterworth Lowpass Filter suitable for use with the system of Fig. 15;
  • FIG. 22 is a block diagram illustrating tone detectors and channel selection for the system of Fig. 15.
  • the radio spectrum is like real estate; they are not making any more of it.
  • the important quantity to conserve is not necessarily the bandwidth itself, but its practical ability to carry as much information as possible.
  • wasting the recoverable information present in strong interfering signals is, quite literally, wasting the channel's capacity for carrying information.
  • the present invention provides new and improved multiaccess communication systems such as cellular telephony, cordless phones, mobile radio, broadcast radio and the like, embodying novel methods and apparatus for extracting a selected signal from a larger family of such signals having some relationship to the selected signal and overlapping predictable electrical characteristics wherein a desired signal to be received is selected from the family of signals, an estimate is made of the magnitude of at least one predetermined electrical parameter in the remaining unselected signals of the family which would normally interfere with the selected signal, and the estimate is used to substantially cancel the unselected signals and facilitate enhanced reception of the desired signal.
  • any signal in the family of signals may be designated as the desired signal to be received to thereby increase system quality, capacity, and availability to an increased number of users.
  • what would normally be interference is selectively harnessed to increase the availability of overlapping signals to multiple users.
  • the estimation of an interfering waveform is used to cancel interference in the form of undesired signals wherein any or all of the signals may be information bearing signals or constructed signals having known and predictable characteristics, as opposed to "white noise" having maximum unpredicability.
  • Such selective interference cancellation yields multiple access to the pool of undesired signals for selective extraction of desired signals by multiple users.
  • the improved system provides receivers which are independent of any power hierarchy in that the desired signal extraction does not depend on relative position of the desired signal in the power hierarchy.
  • the system can detect the weakest signal in the family of overlapping signals, and the relative power hierarchy need not be stable or preserved.
  • the aforedescribed invention may be embodied in various analog and digital receiver configurations, utilizing cross-coupled phase locked loop technology, in configurations embodying inphase, inphase and quadrature, feedforward and difference amplitude-tracking electrical networks.
  • a network system which achieves increased capacity by planned use of interference cancelling receivers to receive each user's desired signal in the presence of interference.
  • a plurality of antennas 101 propagate corresponding signals (Signal 1, Signal 2 ... Signal K) via any suitable network transmission medium 102, as by radio waves, coaxial cable, wires, fiber optics, or the like, characterized by the usual losses, distortion, superposition, etc.
  • the resultant signals are received by an interference cancelling receiver 103, or a plurality of such receivers and demodulated output for a selected signal is available on line 104 for the i th user in a family of multiple users.
  • interference estimates are subtracted from a single input channel.
  • an estimate of the interference is provided as negative feedback, over line 106, from an estimator/demodulator subsystem 105 to a summing junction 107 where it is effectively subtracted from the input to the junction on line 108.
  • the resultant is directed to the demodulation subsystem 105 over line 109.
  • demodulation and estimation subsystems can be either combined or separate processes. Referring now specifically to Fig. 3, wherein like reference numerals denote like or corresponding parts throughout the drawings, there is shown an alternative embodiment wherein the demodulation and estimation processes are separate.
  • a demodulator is only needed for the desired user's signal.
  • the demodulator 105a for the i th user receives input over line 109a from the summing junction 107, while the interference estimator 105b for all users except the i th user, receives such input over line 109b.
  • the negative feedback to the junction 107, over line 106, is provided solely by the estimator 105b.
  • demodulation and estimation are functionally combined, as is the case in our cross-coupled FM interference cancelling receivers.
  • a demodulator/estimator 105c, 105d is needed for each interfering signal.
  • FM frequency-modulation
  • the detected SNR is markedly improved if the input is above a certain received SNR, with the amount and position of the improvement threshold being determined by the bandwidth expansion as given by the ratio of the FM deviation to the modulating signal's bandwidth.
  • this effect known as the FM capture effect, facilitates a simple method of providing multiaccess communications by dividing the service area into geographical regions.
  • the allocation of frequencies for FM transmitters has been traditionally accomplished in this manner.
  • the allowable geographical separation of co-channel and adjacent channel FM transmitters be they high-power broadcast stations or low-power cellular telephones, has been regulated so that propagation laws will mitigate any interference problems.
  • the capture-effect phenomenon of FM reception is exploited to produce simple interference-cancelling receivers with a cross-coupled topology.
  • Phase-locked loop receivers cross-coupled with amplitude-tracking loops to estimate the FM signals are utilized in various embodiments of the invention.
  • the function of these cross-coupled phase-locked loop (CCPLL) interference cancelling receivers is explained below. New interference cancellers inspired by optimal estimation and the CCPLL topology are described, resulting in simpler receivers than those in prior art.
  • FM interference-cancelling receivers are considered for increasing the frequency reuse in a cellular telephone system.
  • Interference mitigation in the cellular environment is seen to require tracking of the desired signal during time intervals when it is not the strongest signal present.
  • Use of interference cancelling in conjunction with dynamic frequency-allocation algorithms is viewed as a way of improving spectrum efficiency. Performance of such interference cancellers indicates possibilities for greatly increased frequency reuse. The known ability of an
  • FM broadcast receiver to unambiguously receive one station or another at the same frequency in areas of interference leads one to conclude that there exist situations where, in accordance with the present invention, both mutually interfering signals could potentially be received with the right receiver.
  • demodulated FM information enhanced by the capture effect, is provided as output over line 104a from a capture effect receiver 100a and first is remodulated as an estimate of the interference. This estimate is then subtracted at junction 112 from the input to a second capture effect receiver 100b, leaving the originally weaker signal now the stronger and allowing its capture by the second receiver.
  • the amplitude component of each interfering signal is estimated, producing an estimate of the entire interfering signal.
  • the full separation of the interfering signals can be exploited. This can be contrasted with tracking only the phase, as is done in conventional broadcast FM receivers.
  • Phase-only tracking can be seen as projecting the resultant of the interfering signals onto the surface of a sphere of constant amplitude in signal space. The loss of sometimes essential side information contained in the resultant's amplitude plagues these receivers.
  • Such "power-division multiplexing" of FM signals has been proposed before, but the separation methods involved tuning a frequency domain "trap" filter, and are, therefore, fundamentally different from the coherent methods described here.
  • the estimation of an interfering waveform is used to cancel interference in the form of undesired signals wherein any or all of the signals may be information bearing signals or constructed signals having known and predictable characteristics, as opposed to "white noise".
  • Such selective interference cancellation yields multiple access to the pool of undesired signals for selective extraction of desired signals by multiple users.
  • Power hierarchy is maintained by direct measurement and tracking, with or without additional power control.
  • successive interference cancelling relies on transmitted power control to maintain a strict power hierarchy.
  • Interference cancelling receivers in accordance with the invention, do not require a strict power hierarchy because they directly measure and track the amplitudes of the received signals. However, power control can still be added on top of this tracking to boost the power of signals that are shadowed from the receiver, with rapidly varying power levels being tracked by the receiver.
  • Such successive interference cancelling receivers fitted with the direct measurement and tracking of power are considered within the scope of the present invention.
  • Figs. 6 and 7 of the drawings there are shown two methods for expanding interference cancelling receivers, in accordance with this invention, to, by way of example, three interfering signals in a family comprising signals S, I1 and I2.
  • independent interference estimators 105d are provided for each of the interference signals from the family of signals.
  • Fig. 6 illustrating only one of the three possible branches of this interference cancelling receiver. The other two branches (not shown) would differ only in that they would have interference estimators for signals I1 and S or I2 and S.
  • a vector sum interference estimator 105e is provided. Again, as in the case of Fig. 6, only one of the three branches is shown for purposes of simplified illustration.
  • a receiver can also estimate interfering signals by a means of phase estimation.
  • the phase-locked loop is not the only way to estimate the phase of an interfering signal.
  • Other phase estimators would produce equivalent systems. For example, systems with frequency-locked loops together with inphase and quadrature amplitude estimation would produce the same result.
  • non-tracking signal phase estimators such as are used in digital phase modulated signal reception could also be used to estimate the phase of the interference.
  • a receiver can also estimate interfering signals by a means of amplitude estimation.
  • the correlators employed in our preferred embodiment are not the only way to estimate the amplitude of an interfering signal.
  • Other amplitude estimators would produce equivalent systems.
  • the several different types of the analog FM interference canceller differ only in the amplitude estimators (feedforward inphase-only, feedforward inphase and quadrature, difference amplitude tracking inphase-only, difference amplitude tracking inphase and quadrature, feedback inphase-only).
  • Other amplitude estimation means such as filtering for a pilot tone and detecting its power, can be used to produce an equivalent system.
  • Still another possible receiver separates interfering signals by means of estimating the propagation loss.
  • Estimation of the amplitudes of the received signals in our interference cancelling receivers is, in fact, a way of estimating the channel propagation loss of each signal. Other methods of channel estimation would be equivalent. Interfering signals would then be separated by identifying the channel loss characteristics associated with each of them.
  • channel estimation includes estimation of the channel impulse response or transfer function, as is done in equalization, and separating interfering signals using the fact that they have passed through different propagation channels. Another method would be estimation of the received signal power and separation of the signals by use of the different fading and shadowing environments. Both of these estimation techniques could be used in a tracking mode, where the estimate is updated over time using previous estimates or, in a non-tracking mode, where new estimates are formed as needed without regard for the previous estimate.
  • Prior art contains many examples of equalization and channel estimation.
  • the channel estimate is not being used only to improve a single desired signal, but also as an identifying characteristic to separate and estimate multiple interfering signals for cancellation. Referring now to Fig.
  • the CCPLL interference canceller shown is an example of subtracting an estimate of interfering signals from the incoming corrupted signal to produce a clean output.
  • the CCPLL configuration includes a junction 113a, phase locked loop 114a and gain control 115a cross-coupled to a junction 113b, phase locked loop 114b and gain control 115b.
  • the estimator-subtractor system is quite general, similar to Fig. 2 previously described, and can be applied to many different interference situations including FM multipath, co-channel FM interference, and AM-FM interference.
  • Getting the best possible estimate of the interference environment is the most important and one of the most challenging aspects in developing one of these interference cancellers.
  • Estimation approaches vary from those that utilize extensive a priori knowledge of the input to those that make as few restrictions as possible on the input signals.
  • the CCPLL interference canceller is based on the topology of a noncausal optimum (Maximum a Posteriori (MAP) ) FM receiver for suppression of interchannel interference with constant modulus (no AM component). Given an input signal mix:
  • MAP Maximum a Posteriori
  • s 1 (t) A 1 cos( ⁇ 1 t - x 1 (t))
  • s 2 (t) A 2 cos( ⁇ 2 t + x 2 (t))
  • n (t ) is white Gaussian noise with zero mean and normalized two-sided power spectral density of 1 W/Hz.
  • x 1 (t) is the desired-signal angle modulation (here assumed a Gaussian process), for FM
  • x 1 (t) ⁇ m 1 (u ⁇ du
  • m 1 (t) is the desired-signal frequency modulation information.
  • x 2 (t) is the interfering signal angle modulation (also assumed Gaussian).
  • a 1 and A 2 are the moduli of the two signals and are assumed constant.
  • k 1.2 constants
  • R x1,2 (t,u) are the impulse responses of lowpass filters, and represent the minimum mean- square error (MMSE) estimates of s 1 (t) and s 2 (t).
  • MMSE minimum mean- square error
  • the CCPLL system utilizes two important characteristics of the phase-locked loop (PLL) to produce its estimate:
  • a PLL is an unbiased estimator of phase.
  • PLL's like other FM demodulators, exhibit FM capture effect and nonlinearly enhance the strongest signal when more than one signal is present.
  • the characteristics (loop bandwidths, gain control, etc.) of the CCPLL interference canceller can be fine-tuned to best accommodate different types of interferers.
  • this requires a priori knowledge of both the input desired signal and the interfering signals.
  • Modifications of the system to include amplitude-tracking loops inspired by Least Mean Squares (LMS) adaptive filters have been reported to produce a slightly more complex system that is fairly robust in the presence of co-channel AM-FM and pulsed RF (FM) interferers.
  • LMS Least Mean Squares
  • FM pulsed RF
  • LMS-type amplitude-tracking loops transforms the device from one based primarily on the FM capture effect into a more standard, adaptive noise-cancelling device.
  • adaptive noise cancelling topology often called the "adaptive notch filter”
  • a periodic reference is used in both inphase and quadrature (90 degrees phase-shifted) forms to cancel correlated components in the input.
  • the LMS algorithm is used to produce the inphase and quadrature multiplicative weights necessary to cancel the reference's component from the input.
  • the implementation of the CCPLL interference canceller with amplitude control may also be in the form so-called Vector-locked Loops (not shown).
  • the resulting implementation is a so-called self-bootstrapping, continuous analog version of the discrete digital adaptive filter, with the reference inputs being provided by the VCO's in the capture effect PLL's.
  • the system is called self-bootstrapping because it can acquire the interfering signals by itself because of the FM capture effect.
  • Experimental trials with the CCPLL system have reported that acquisition is a two-stage process, as might be expected. Software simulations have been reported, confirming this finding. The first stage consists of a PLL capturing and locking to the strongest signal present. The second phase is marked by the convergence of the amplitude control weights. The two phases are then repeated by the second PLL, resulting in FM demodulation of both the weaker and the stronger signal.
  • the inputs to the amplitude estimators often contain some leakage from the undesired signal. This cross leakage affects the convergence of the amplitude-control loops and results in further imperfections in cancellation. Because the CCPLL structure contains a reference for each signal present, it is plausible to think of recursively applying the adaptive noise canceller in an attempt to limit cross leakage. A successful second application of the adaptive noise canceller, cross-coupling the VCO references themselves, would serve to decrease components of the undesired signal in each reference and, therefore, be potentially more robust.
  • PLL #2 The effect of a residual, strong signal component in the VCO output of the weak signal tracking PLL 114b (referred to as PLL #2) on its amplitude-control loop 115b can be considered as leakage of signal into the reference input of an adaptive notch filter. This will cause some cancellation of the strong signal prior to the input of the strong signal tracking PLL 114a (PLL #1).
  • the strong signal is denoted Si and the weak signal s 2 ; the transfer function for the weak signal to the VCO of PLL #2 is H 2,2 (f).
  • S x2x2 (f) S s2s2 (f)
  • the FM capture effect further remforces , and provides the feedback necessary to
  • the system shown in Fig. 11 adds "leakage-control loops" 120a and 120b to the VCO outputs of the two PLL's.
  • These adaptive noise-cancelling loops attempt to better orthogonalize the two VCO references, increasing capture in the interference-cancelling reference.
  • the use of a second pair of adaptive noise-cancelling loops would cause the SIR at the input of the PLL's to converge to the inverse square of the SIR produced by the PLL's, rather than just to the inverse of the SIR at the PLL output.
  • the resulting system has decreased acquisition time.
  • leakage-control loops 120a, 120b in Fig. 11 creates a stable, steady-state solution where both PLL's are tracking the stronger signal. This new stable state is equivalent to an uncoupled system. While forcing the VCO outputs to be uncorrelated with respect to each other, the leakage-control loops produce the degenerate solution where both VCO outputs are uncorrelated to either signal. This, in turn, causes little or no cancellation of the strong signal at the input to either VCO, and strong-signal capture occurs in both PLL's, reinforcing the stable state.
  • the leakage-control loops configuration is not used in the preferred embodiment.
  • a CCPLL-based interference-cancelling receiver is to be extended to more than two signals, simply tracking the amplitude of each received signal is not necessarily optimal. This is because a minimum mean-square error estimate for the sum total interference is desired, and the MMSE estimate of a sum is not necessarily the sum of the individual MMSE estimates.
  • interference amplitude-tracking loops are associated with each desired signal rather than with each interfering signal. The amplitudes of the interfering signal components to be subtracted from the input are estimated jointly for each desired signal in method paralleling multireference adaptive interference-cancellation techniques. If the VCO-produced references of the summed interferers are uncorrelated with each other, the estimation of the amplitudes of the summed interference will be the same as separately estimating the amplitudes of the component signals.
  • Amplitude estimation for only the inphase portion of each signal should suffice, in some instances, reducing the hardware of the CCPLL dramatically.
  • Implementation of this system can be attempted either with a feedforward amplitude estimator using the form of the estimator shown in Fig. 12 of the drawings, or in a feedback loop form shown in Fig. 13 by using the gradient-descent form of the optimal solution.
  • the amplitude estimator in Fig. 12 adds a pair of lowpass filters 121a, 121b in a direct form, feedforward optimum amplitude estimator wherein the input and phase estimate are combined to produce an amplitude estimate of the interference, and where the amplitude estimator does not form a loop including the actual cancellation or subtractor operator.
  • the amplitude estimator in Fig. 13 also uses lowpass filters 121a, 121b, but in a gradient-descent form, feedback optimum amplitude estimator wherein the input interference is cancelled by the previous estimate of interference.
  • the residual is correlated against the phase estimate of interference to produce the next iteration or update of the amplitude interference estimate.
  • the direct form shown in Fig. 12 simplifies the dynamics of the system by replacing a feedback loop by a feedforward loop.
  • the weight estimate is just the lowpass portion of the product of the original input signal (without any cancellation) and the inphase estimate.
  • the lowpass filters 121a, 121b are to have a cutoff frequency high enough to pass the amplitude information, which is typically near the lower band edge of the modulating information.
  • the estimator is a coherent AM receiver with the PLL's coherent phase reference as its coherent carrier reference. The coherent AM receiver thus tracks the amplitude of the FM signal just the same as if it were modulated on a constant-frequency carrier.
  • the gradient-descent solution has the form:
  • E[e 2 ] denotes the expected value over the ensemble of noise.
  • the gradient-descent solution in Fig. 13 is just the inphase part of the amplitude-tracking loop given in the iterative CCPLL.
  • the gradient-descent implementation form requires integrators 122a, 122b which the direct form does not require.
  • v(t) A 1 cos( ⁇ c t + ⁇ 1 ) + A 2 cos( ⁇ c t + ⁇ 2 ) .
  • the first four equations specify the amplitude-control loops, while the last two specify the phase-locked loops. All the variables are real, with the amplitude estimates being represented by x (1,2)(I,Q) and the phase estimates by ⁇ 1,2 .
  • the symbol * denotes convolution.
  • the loop filters for PLL's 1 and 2 are, respectively, h L1 and h L2 , and the lowpass filters in the amplitude-control loops are g 1I , g 1Q , g 2I , and g 2Q .
  • the PLL gains are ⁇ q and ⁇ 2
  • the amplitudecontrol loop gains are ⁇ 1 and ⁇ 2 .
  • limiters would have a significant effect during signal acquisitions by enhancing the strong and suppressing the weak signals. Since the amplitude-control loops are completely preceding the limiter, the limiter does not affect their operation except through its effect on the phase estimates. As a result, in a system with bandpass limiters, the first four differential equations above are unchanged. However, since the limiter nonlinearly mixes the additive signal components before they reach the PLL, the operations of the PLL cannot be distributed among the additive signal components as before. This is because the limiter has already distorted the linear combination of these components before they get to the PLL.
  • ⁇ 1 ⁇ 1 [sin( ⁇ 1 )BPLIM(A 1 cos ⁇ 1 + A 2 cos ⁇ 2 -x 2/ cos ⁇ 2 -x 2Q cos ⁇ 2 )]*h L1
  • ⁇ 1 ⁇ 1[sin( ⁇ 2 )BPLIM (A 1 cos ⁇ 1 + A 2 cos ⁇ 2 -x 1/ cos ⁇ 1 -x 1Q cos ⁇ 1 )]*h L2
  • BPLIM(y) denotes the bandpass limiter's operation on
  • the gradient-descent form of the reduced solution, derived earlier and shown in Fig. 13 follows by noting that its only difference from the system defined by the six equations above is that the guadrature amplitude estimates are set identically zero.
  • amplitude estimates x 1 and x 2 are computed directly as follows:
  • x 2 [A 1 cos( ⁇ 1 - ⁇ 2 ) + A 2 cos( ⁇ 2 - ⁇ 2 )]*g 2 .
  • ⁇ 1 ⁇ 1 [A 1 sin( ⁇ 1 - ⁇ 1 ) + A 2 sin( ⁇ 2 - ⁇ 1 ) - [(A 1 cos( ⁇ 1 - ⁇ 2 ) + A cos( ⁇ 2 - ⁇ 2 )) *g 2 ]sin( ⁇ 2 - ⁇ 1 )] * h L1
  • ⁇ 2 ⁇ 2 lA 1 sin( ⁇ 1 - ⁇ 2 ) + A 2 sin( ⁇ 2 - ⁇ 2 ) - [(A 1 cos( ⁇ 1 - ⁇ 1 ) + A 2 cos( ⁇ 1 - ⁇ 1 )) *g 1 ]sin( ⁇ 1 - ⁇ 2 )] * h L2 .
  • the loops will track the input phase rate (instantaneous frequency) as well as the input phase.
  • We will consider a perturbation to the phase variable and consider the average phase rates tracked by the loops in order to remain in lock. In real systems this approach is further justified by the time constants associated with the loop lowpass filter, which are noted for the ability to "flywheel" the loop through relatively short noise bursts. Therefore, let the loop initial conditions be:
  • ⁇ 1 ⁇ 1 [A 1 sin(- ⁇ ) + A 2 sin( ⁇ t + ⁇ - ⁇ ) - [[A 1 cos(- ⁇ t- ⁇ ) + A 2 cos( ⁇ - ⁇ )]*g 2 ]sin( ⁇ t + ⁇ - ⁇ )]*h L1
  • ⁇ 2 ⁇ 1 [A 1 sin(- ⁇ t- ⁇ ) + A 2 sin( ⁇ - ⁇ ) - [[A 1 cos(- ⁇ ) + A 2 cos( ⁇ t + ⁇ - ⁇ )]*g 1 ]sin(- ⁇ t- ⁇ + ⁇ )]*h L2
  • ⁇ ⁇ 2 - ⁇ 1 .
  • g 1 and g 2 be narrowband filters so that they cut off below frequency ⁇ , as they would be for tracking amplitude functions whose bandwidths are less than those of the modulating information on an FM signal.
  • ⁇ 2 ⁇ 1 [A 1 sin(- ⁇ t- ⁇ ) + A 2 sin( ⁇ - ⁇ ) -A 1 cos(- ⁇ )sin(- ⁇ t - ⁇ + ⁇ )]*h L2 . If ⁇ is outside the pass band of the lowpass loop filters h L1 and h L2 , the system displays the dynamics of ordinary, first-order phase locked loops without any coupling. More importantly, the system behaves as if the interference were not present:
  • ⁇ 2 ⁇ 2 A 2 sin( ⁇ - ⁇ ) .
  • ⁇ 1 ⁇ 1 A 1 ( ⁇ 1 - ⁇ 1 )
  • ⁇ 2 ⁇ 2 A 2 ( ⁇ 2 - ⁇ 2 ) .
  • ⁇ 1 ⁇ 1 [A 1 sin(- ⁇ ) + A 2 sin( ⁇ t + ⁇ - ⁇ ) - [[A 1 cos ( ⁇ ) + A 2 cos(( ⁇ t+ ⁇ - ⁇ )]*g 2 )sin( ⁇ - ⁇ )]*h L1
  • ⁇ 2 ⁇ 2 [A 1 sin(- ⁇ ) + A 2 sin( ⁇ t + ⁇ - ⁇ ) - [[A 1 cos(- ⁇ ) + A 2 cos( ⁇ t + ⁇ - ⁇ )]*g 1 ]sin( ⁇ - ⁇ )]*h L2
  • ⁇ 1 ⁇ 1 [A 1 sin(- ⁇ ) - A 1 cos(- ⁇ )sin( ⁇ - ⁇ )]
  • ⁇ 2 ⁇ 2 [A 1 sin(- ⁇ ) - A 1cos(- ⁇ )sin( ⁇ - ⁇ )] .
  • ⁇ 1 G ( ⁇ 1 - ⁇ 2 )
  • ⁇ 2 G ( ⁇ 1 - ⁇ 1 ) .
  • phase error for ⁇ 1 , ⁇ will cross over zero while the phase error for ⁇ 2 , ⁇ , is still on the original side of zero, providing us the perturbation case where ⁇ and 7 are of different signs.
  • ⁇ 1 ⁇ i[A 1 sin(- ⁇ )+ A 2 sin( ⁇ t + ⁇ - ⁇ ) ]*h L1
  • ⁇ 2 ⁇ 2 [A 1 sin(- ⁇ )+A 2 sin( ⁇ t + ⁇ - ⁇ ) ]*h L2 .
  • ⁇ and ⁇ are time-varying functions. These are also the equations governing the two phase-locked loops without any cross-coupling. At this point the loops begin to reacquire the signal as normal PLL's would since the entire assembly has lost lock on both the amplitude and the phase of either signal. There is one major difference, however, in that one of the loops' frequency estimates is biased toward the interfering signal, and one is biased away from it. The loop whose frequency is moving towards ⁇ 2 will become less effective in filtering out that signal's power, and as a result, will be significantly slower in reacquiring the first signal. The cancellation of the first signal at this loop's input will have time to take hold, allowing the possible acquisition of the second signal.
  • the amplitude estimate will be corrupted by the undesired signal.
  • the undesired signal will have a corrupting effect, somewhat diminished, even when the instantaneous frequency difference is slightly outside the nominal passband of these filters.
  • This unavoidable corruption of the amplitude estimate can be viewed as an undesired signal-leakage path into the interference-canceller reference, an effect that limits performance of the canceller discussed. It is, therefore, desirable to suppress any potential cross-leakage in generating the amplitude estimate.
  • One possible way of limiting such cross-leakage is to derive the amplitude estimate from as clean a source of the desired signal as it possible.
  • the canceller topology already contains such a source during steady-state operation.
  • the input to each limiter-PLL assembly has already had the estimate of the undesired signal subtracted from it, and it is a good source from which to produce the desired-signal amplitude estimate.
  • computing the amplitude estimate one could multiply the VCO-produced signal estimate not by the overall canceller input signal v(t), but rather by the input to its own limiter-PLL combination. The effect of leakage can be minimized in the canceller steady state by doing so.
  • This canceller topology is shown in Figure 14 and is referred to as the difference-amplitude-tracking topology.
  • the difference- amplitude tracking technique where the amplitude is tracked after the interference has been received, is a novel and particularly useful subsystem. Since this is used to estimate amplitude variation, either caused by channel distortion or deliberate variation, it is separable from the FM interference canceller as a whole.
  • the difference amplitude tracking topology can be formed as Inphase-only and Inphase-and-Quadrature, and can be applied to any interference cancelling receiver, not only FM.
  • a 2 [A 1 cos( ⁇ 1 - ⁇ 2 ) + A 2 cos( ⁇ 2 - ⁇ 2 ) - a 2 cos( ⁇ 1 - ⁇ 2 )] * g 2 .
  • ⁇ 1 ⁇ 1 [A 1 sin(- ⁇ ) + A 2 sin( ⁇ t + ⁇ - ⁇ ) - a 2 sin( ⁇ t + ⁇ - ⁇ )]*h L1
  • a 1 [A 1 cos(- ⁇ ) + A 2 cos( ⁇ t- ⁇ ) - a 2 cos( ⁇ t + ⁇ - ⁇ )] * g 1
  • a 2 [(A 1 -a 1 cos( ⁇ )+a 1 sin( ⁇ ))cos(- ⁇ t- ⁇ ) + A 2 cos(- ⁇ )] * g 2 .
  • any leakage of the undesired signal through the amplitude lowpass filter will be significantly attenuated for small ⁇ and ⁇ .
  • the perturbation is reduced to the same system referred to earlier after the amplitude filtering:
  • ⁇ 1 ⁇ 1 [A 1 sin(- ⁇ ) + A 2 sin( ⁇ t + ⁇ - ⁇ ) - A 2 cos( ⁇ - ⁇ )sin( ⁇ t + ⁇ - ⁇ )]*h L1
  • ⁇ 2 ⁇ 2 [A 1 sin(- ⁇ t- ⁇ ) + A 2 sin( ⁇ - ⁇ ) - A 1 cos(- ⁇ )sin(- ⁇ t- ⁇ + ⁇ )]*h L2 .
  • a 1 [A 1 cos(- ⁇ ) + A 2 cos( ⁇ t + ⁇ - ⁇ ) - a 2 cos( ⁇ - ⁇ )] * g 1
  • the two amplitude estimates are underdetermined and, hence, they can vary widely in relation to the signal amplitude, A 1 , while still satisfying the defining equation.
  • the amplitude estimates are free to wander, possibly producing unstable behavior. In a practical system this effect would be dealt with by disabling the canceller when no significant interference is present, as indicated by the residual power level after the desired signal has been removed.
  • ⁇ 1 ⁇ 1 [A 1 sin(- ⁇ ) + A 2 sin( ⁇ t + ⁇ - ⁇ ) - a 2 sin( ⁇ - ⁇ )]*h L1
  • ⁇ 2 ⁇ 2 [A 1 sin(- ⁇ ) + A 2 sin( ⁇ t + ⁇ - ⁇ ) - a 1 sin(- ⁇ + ⁇ )]*h L2 .
  • ⁇ 1 ⁇ 1 [A 1 sin(- ⁇ ) - (A 1 -a 1 )sin( ⁇ - ⁇ )]
  • ⁇ 2 ⁇ 2 [A 1 sin(- ⁇ )+ ⁇ 1 sin( ⁇ - ⁇ )], which, for small ⁇ , y linearizes to:
  • ⁇ 1 - ⁇ 1
  • ⁇ 2 - ⁇ 2 [a 2 ⁇ + a 1 ⁇ ] .
  • Cross-coupled phase-locked loops (CCPLL) -based FM interference-cancelling receivers are suitable as an interference canceller for use in multiaccess communications.
  • CCPLL interference canceller Based on an understanding of the operational principles behind the CCPLL interference canceller and on optimal estimation theory, new interference-cancelling systems are provided in accordance with the invention. Some of the new systems are relatively simple and provide only one amplitude-tracking loop. Others have improved steady-state operation, obtained by forming the estimate of a signal's amplitude from an interference-cancelled reference.
  • the parameters of these receivers i.e., the loop bandwidth, the loop filter transfer function, and the amplitude-estimation filters
  • the parameters of these receivers can be optimized by a known technique, i.e., matched filtering.
  • a known technique i.e., matched filtering.
  • interference cancellers demonstrate that reception of both of two co-channel FM interferers is possible with such cancellers.
  • a feedforward amplitude-estimation procedure results from direct solution of the MMSE amplitude equations. This simplifies the resulting system dynamics by reducing the number of defining differential equations from six to a maximum of four for both inphase and quadrature tracking.
  • the tracking of the amplitude as a complex value is shown to provide little cancellation gain when the PLL generally tracks close in phase.
  • the importance of the quadrature component thus depends on the reception environment and the PLL tracking parameters.
  • the tracking loop for the quadrature-amplitude component not only increases the hardware requirements but also increases the amplitude-tracker noise bandwidth.
  • the difference-amplitude-tracking topology has more complicated dynamics than the simpler feedforward cancellers.
  • successful separation and demodulation of each of two co-channel interfering FM modulated signals is achieved by all four feedforward interference-canceller topologies (inphase and quadrature amplitude-tracking, inphase-only amplitude tracking, inphase and quadrature difference-amplitude-tracking, and in-phase-only difference-amplitude-tracking).
  • Such interference cancellers may be specifically applied to the cellular telephone channel.
  • the cellular telephone channel is an important example of a multiuser channel since it uses the natural geographic attenuation of signals to allow a limited number of frequency slots to be reused in cells separated in the service area.
  • the cellular channel Since cellular voice transmission is full duplex as opposed to push-to-talk, the cellular channel is naturally divided into mobile-to-base and base-to-mobile frequency sets. Because of the requirement for having competing cellular systems available in each metropolitan area, the available bandwidth in each set is then halved between two competing carriers.
  • the individual channel spacing in the United States is set at 30 kHz, and voice transmission is FM with pre-emphasis and a maximum peak frequency of deviation of ⁇ 12 kHz.
  • Propagation in the individual channels has been modeled as an inverse fourth-order law with Rayleigh fading and large scale average lognormal shadowing. Since the shadowing is highly geographically dependent and occurs in an average (long-term) signal level sense, it will not be dealt with further here.
  • the Rayleigh fading devastates the on-average capture effect. Fades of from 10 to 40 dB with durations on the order of one to ten milli-seconds occur at rates from 1 to 100 per second. However, it should be remembered that capture can be observed on the time scale of the FM channel bandwidth, or at least the tracking loop bandwidth for PLL's, a minimum of ⁇ 3 kHz. As a result, demodulator outputs can and do actually show signal capture during the relatively long time between fades.
  • the cellular channel has 24 kHz allocated for voice transmission.
  • Voice is syllabically companded (companded with a time constant on the order of 20 milli-seconds), pre-emphasized and frequency-modulated.
  • the peak frequency deviation is 12 kHz, maintained by amplitude clipping if necessary.
  • the rms frequency deviation is 2 kHz, with f 2 rms log normally distributed with standard deviation of about 5 dB.
  • the current FM cellular telephone receiver is a simple limiter-discriminator demodulator. While the limiter-discriminator based receiver is inexpensive and simple to produce, advances in analog integrated circuits continually make more complex receiver designs practical. Entire loop-based FM receivers integrated on a chip can be purchased by consumers for under five dollars. At this cost, the cellular telephone receiver, the main element responsible for call quality in a cellular phone costing several hundred dollars or more, can certainly afford to have some money spent on improved performance, especially to increase the traffic capacity.
  • the FM signal must be strong enough to overcome the more frequent fading, and must be fairly robust in the presence of deeper fades.
  • the fading situation is a problem that can be solved by appropriate burst-error-correction codes.
  • Discarding amplitude is an information-lossy process.
  • an amplitude track can at least provide a figure of confidence for the frequency estimates taken at the same time to be used as side information in further filtering processes.
  • the amplitude tracks can be used to feed a received-signal environment model in an effort to separate the signal of interest.
  • the angle information will be corrupted the most when the received amplitude is near its mean, and the received information will be the least corrupted when the amplitude is nearest its maximum and minimum, except in the case of origin encirclements in the phase plane.
  • the basis for the SIR requirement is derived not from considering a signal of constant power, but from the maximum tolerable losses under multipath fading.
  • tolerable variations in maximum acceptable probability of loss that is due to multipath fades vary from .1 as the highest acceptable value for speech down to 0.01 for a high-quality speech system.
  • the problem for the interference canceller is then to reduce the level of interference sufficiently so that track of the desired signal can be maintained through a fade without loss.
  • interference cancellers in such a way as to make use of the difference in instantaneous frequency between uncorrelated, co-channel interfering FM signals.
  • the application of interference cancellers, in accordance with the invention, is directed at receiving both the stronger and the weaker of two signals.
  • the interference canceller must be able to track signals with time-varying amplitudes.
  • the amplitude-estimation filters are chosen to be first-order Butterworth filters with cutoff frequencies larger than the bandwidth of the amplitude modulation.
  • the amplitude-estimation filter is made wider on one loop than on the other.
  • the amplitude loop bandwidths are 400 Hz for the weak-signal estimator and 200 Hz for the strong-signal estimator.
  • the PLL's must be able to separate signals with the same general modulation characteristics, but with widely varying instantaneous power levels.
  • the loop parameters would be proportional to the power ratio.
  • the fading channel makes determination of the exact power ratio to be used difficult, but it is reasonable to design the loops for the mean power ratio.
  • the mean power ratio could be estimated or controlled in order to get the best response, but a fixed value can be chosen.
  • the loops are designed for a mean power ratio close to one.
  • the PLL's are designed using the rules for 3 kHz modulation bandwidth at 12 kHz peak frequency deviation and a weak-to-strong amplitude ratio of 0.85 (1.4 dB SIR, strong to weak). Experimental results have been reported indicating that performance is improved by matching the power level to the "weak" signal loop parameters. As a result, performance may suffer slightly at signal-to-interference ratios away from 0 dB.
  • One advantage of designing around a power ratio close to one is that the bandwidths of both loops are kept small. This permits cleaner phase tracking of both signals in any case, since both PLL's now see a smaller instantaneous bandwidth. Both the fading, which shifts the signal phase as well as its amplitude, and the small loop bandwidths required for operation at a mean power ratio close to one predictably make the instantaneous loop phase errors larger. The error will degrade the performance of the inphase-only amplitude-tracking interference cancellers relative to the inphase and quadrature tracking forms, which can correct for loop phase error in the amplitude estimates.
  • the difference-amplitude-tracking configuration performs better than the feedforward configurations. This is because the fading channel requires amplitude tracking at a bandwidth close to the lower edge of the modulation bandpass. More interfering signal power can come through the amplitude filters, causing corruption of the amplitude estimate.
  • the main advantage of the difference-amplitude configurations is the reduced corruption of amplitude estimates by the interference. In a system where the amplitude-tracking filters are wider, this factor should become important in determining performance.
  • Appendix A which is hereby incorporated herein by reference, which is a document setting forth in greater detail applications of frequency modulation interference cancellers to multiaccess communication systems prepared by and reflecting efforts by the inventor in the development of the present invention. While Appendix A is not considered necessary for a proper understanding to enable one of ordinary skill in the art to make and practice the invention, it is provided for added convenience and insight.
  • Figs. 15 through 22 of the drawings there is shown a presently preferred embodiment of the invention, an interference cancelling receiver system for analog FM transmission similar to that used for cellular telephony in the United States.
  • the preferred embodiment is shown as a receiver for two interfering signals.
  • the number of interfering signals is a parameter in the design, i.e., the receiver design can be directly expanded to cover more than two interfering signals.
  • the preferred embodiment of the invention is a digital hardware implementation of the inphase and quadrature difference amplitude tracking receiver for two interfering FM signals and, as such, is similar to the embodiment of Fig. 14 previously described.
  • the system operates on a sample clock frequency of 120 kHz, with inphase and quadrature sampling.
  • the fundamental word-length in the digital system is 16 bits.
  • the interfering signals and noise enter the system through a conventional antenna (not shown), over line 100, and are conventionally downconverted and filtered to a convenient, standard intermediate frequency (IF) by subsystem 101.
  • IF intermediate frequency
  • the IF is then converted to baseband inphase and quadrature (90 degree phase-shifted) (I&Q) 12-bit digital samples, oversampled at a 120 kHz sample rate by subsystem 102 in Fig. 15. As shown in Fig. 16, the I&Q analog-to-digital
  • A/D conversion subsystem 102 may be of conventional design, formed by splitting the IF on line 110 into two streams through a power divider 111 and mixing one stream at a mixer 112 with cos(2 ⁇ f Ip t), a local oscillator at zero degrees phase, and the other stream at a mixer 13 with -sin (2 ⁇ f Ip t), a local oscillator of 90 degrees phase, where f Ip is the IF frequency.
  • These two signals are then low-pass filtered by a pair of filters 114, 115, respectively, to a 30 kHz bandwidth to produce the desired baseband I&Q signal streams.
  • the baseband I&Q signals are then sampled and quantized by 12 bit A/D converters 116, 117, respectively.
  • the 12 bit digital data streams then enter the interference cancelling FM demodulator subsystem 103, over lines 118, 119, as shown in Fig. 15.
  • the interference cancelling FM demodulator subsystem 103 is shown in great detail in Fig. 17.
  • Digital I&Q samples are input to the FM demodulator subsystem 103, and estimates of the I&Q samples of each of the two interfering signals are subtracted from the input at 120a and 120b, respectively.
  • the resulting difference signals have reduced interference levels, with one signal stronger on the "a" I&Q paths and the other signal stronger on the "b" I&Q paths.
  • Each path pair is input to a phase-locked loop (PLL) of conventional design, comprised of a phase detector 121, loop gain and filter 122, and a numerically controlled oscillator (NCO) 123.
  • PLL phase-locked loop
  • the PLL demodulates the FM signal (at 3a or 3b) and produces an amplitudenormalized estimate (at 4a or 4b) of the signal it is demodulating.
  • This estimate output from the PLL's NCO, enters an amplitude estimator 124 (at 5a or 5b) which constructs the interference estimate (at 6a or 6b) to be subtracted from the signal input of the other signal's demodulator at 120b or 120a.
  • the I&Q samples of the difference signal first enter the phase detector 121 which produces a signal equal to the sine of the phase difference (at 2a or 2b) between the difference signal (at 1a or 1b) and the NCO's current signal estimate (at 4a or 4b).
  • the phase detector 122 may be of conventional design. It first produces an estimate of the instantaneous signal power by squaring each component (I&Q) at (1) and forming the sum I 2 +Q 2 . The reciprocal square root of this value (1/ ⁇ (I 2 +Q 2 )) is then looked up in a table at (2), and multiplied by the I&Q components to normalize the instantaneous magnitude at (3), producing the signals I1 and Q1.
  • the NCO outputs which are produced by the NCO already in normalized form, are input as Inphase Input #2 (I2) and Quadrature Input #2 (Q2).
  • the phase difference output (at 2a or 2b) in Fig. 17 is input to the loop filter and gain 222, shown detailed in Fig. 19, at (1).
  • Loop gain, G is applied by a multiplier 228 whose output is sent to both an integrator 229 at (2) with gain K, and a summer 232 to produce an integral-plus-proportional-type output at (3).
  • This output represents the demodulated FM signal and it is both sent back to the NCO 223 (Fig. 17) to increment the phase value in the NCO, and, as described earlier, out of the FM demodulator (at 3a or 3b in Fig. 17) as the demodulated output.
  • the NCO 223 accumulates the phase increment values input to it, with the resulting sum being an estimate of the signal phase at a given instant in time. It outputs I&Q values representing the cosine and sine of the accumulated phase, respectively.
  • the I&Q output pair (at 4a or 4b in Fig. 17) is used in the phase detector 221 of the PLL, as previously described.
  • the I&Q outputs from the NCO 223 are also input to an amplitude estimator 224 (at 5a or 5b in Fig. 17) where estimates (at 6a or 6b in Fig. 17) of the I&Q input samples of each signal are produced to be subtracted from the opposing PLL's input (at 1b or 1a in Fig. 17).
  • the Inphase and Quadrature Difference Amplitude Tracking form this is accomplished by complex correlation of the NCO outputs (at 4a or 4b in Fig. 17) with the difference, interference-cancelled signal (at 1a or 1b in Fig. 17).
  • the amplitude estimator 224 is shown in greater detail in Fig.
  • the sample estimator can be thought of in two stages.
  • the first stage is the production of I&Q narrowband amplitude estimates by complex correlation.
  • the second stage is the complex modulation of I&Q outputs from the NCO 223 (Fig. 17) by the amplitude estimates to produce the I&Q sample values.
  • the complex correlation is performed as a complex conjugate multiplication followed by lowpass filtering by lowpass filters 230.
  • the I&Q difference samples (at 1a or 1b in Fig. 17) are input as 1I and 1Q in Fig. 20, and the I&Q NCO outputs (at 4a or 4b in Fig.
  • 3I and 3Q are then each lowpass filtered to form the I&Q amplitude estimates 4I and 4Q at (3).
  • the lowpass filter 230 used in Fig. 20 follows a basic first-order recursive design.
  • the coefficients are for a first-order digital Butterworth filter.
  • the same basic filter design, with different bandwidths (controlled by the filter coefficients) is used in all the digital lowpass filters in this embodiment of the invention.
  • Figure 21 of the drawings illustrated the lowpass filter 230 in greater detail.
  • the signal samples to be filtered are input at (1).
  • the input sample, and a one-sample delayed version of the input are each multiplied by their respective coefficients, CI1 and CI2 respectively, and are summed at (2). This sum is then added to the product of the previous output and its coefficient, CO1 (3), producing the new output at (4).
  • This new output is both output from the filter 230 and placed in the one-sample delay register to be used in the next sample's computation. Note, that, for the lowpass filter 230 shown in more detail in Fig. 21, if any coefficient (CI1, CI2, or CO1) is 1.0, the multiply associated with that coefficient is removed from the filter.
  • the multiplier becomes a shifting operation, e.g., a multiply by 4 is a logical left shift of 2 bits.
  • Negative powers of two e.g., 1/4, are logical right shifts. Negative coefficients can be handled as above, with the following addition changed to a subtraction.
  • Cancelling FM Demodulator 203 are passed to the Tone Detector and Channel Selection subsystem 204 in Fig. 15.
  • Fig. 22 illustrates the latter subsystem 204 in greater detail.
  • Both demodulated inputs (X1 and X2) are multiplied by the I&Q outputs of an NCO (YI and YQ at (2)), producing the I&Q product pairs (YI*X1, YQ*X1), and (YI*X2), YQ*X2).
  • Each signal is passed through a lowpass filter 230, which must be narrower than the differences of the two signal marker frequencies.
  • the filtered outputs at (3) are Z1I, Z1Q, and Z2I, Z2Q.
  • each I and Q pair are summed to form the signal powers as in the phase detector, i.e., Z1I*Z1I + Z1Q*Z1Q, and Z2I*Z2I + Z2Q*Z2Q.
  • the power values are compared by a comparator 233 at (4), resulting in the selection (5) at multiplexor 234 of the demodulated output with the strongest component of the desired marker frequency for output at (6).
  • the resulting selected output is then conventionally converted via a digital to analog converter 205 and lowpass filtered through a conventional analog filter 206 to the desired signal bandwidth at (5) in Fig. 15, to provide the desired demodulated and filtered output over line 207.
  • Circuit Parameters for the system of Figs. 15 through 22 are as follows:
  • PLL Phase-Locked Loop
  • phase detector gain is 1 unit/rad.
  • Loop Gain G1 (PLL #1 (a)) or G2 (PLL
  • G1 and G2 depend on the NCO's chosen.
  • CO1 (1.-tan( ⁇ /600))/(1.+tan( ⁇ /600)) Marker frequencies should be between 4 kHz and 5 kHz, separated by at least 400 Hz.
  • the new and improved multiaccess communication system of the present invention is unique in that it harnesses interference to increase system capacity available for plural access by single or multiple users.
  • the improved system provides receivers which are independent of any power hierarchy in that desired signal extraction does not depend on relative position of the desired signal in the power hierarchy.
  • the system can detect even the weakest signal in the family of overlapping signals, and the relative power hierarchy need not be stable or preserved.
  • the system of the present invention satisfies a long existing need in the communications industry for an accurate, reliable, easy to use, relatively simple and economical system for exploiting the frequency spectrum to provide increased system capacity available to multiple users.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Noise Elimination (AREA)

Abstract

Un système de communications à accès multiple pour téléphones cellulaires, téléphones portatifs, communications radiophoniques mobiles, émissions de radio et de télévision, réseaux de communications radiophoniques et similaires, comprend l'utilisation planifiée de procédés et d'appareils de réception à annulation d'interférences qui permettent à des utilisateurs multiples de partager les mêmes fréquences attribuées, ou des fréquences attribuées qui se chevauchent, par évaluation des signaux non voulus dans une famille de tels signaux, de manière à les annuler et à recevoir les signaux voulus sélectionnés dans ladite famille de signaux. L'invention concerne plusieurs récepteurs à boucle à verrouillage de phase à couplage diaphonique, aussi bien analogiques que numériques, et qui utilisent des configurations de poursuite en phase, en phase et en quadrature, à action directe et à amplitude différentille. Ces récepteurs peuvent être utilisés, à titre d'exemple, pour mettre en ÷uvre l'invention.
PCT/US1993/006125 1992-06-26 1993-06-25 Procede et appareil d'annulation d'interferences dans un recepteur de trafic a acces multiple WO1994000918A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
AU46533/93A AU4653393A (en) 1992-06-26 1993-06-25 Method and apparatus for interference cancellation in a multiple access communication receiver

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US91082992A 1992-06-26 1992-06-26
US07/910,829 1992-06-26

Publications (1)

Publication Number Publication Date
WO1994000918A1 true WO1994000918A1 (fr) 1994-01-06

Family

ID=25429381

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US1993/006125 WO1994000918A1 (fr) 1992-06-26 1993-06-25 Procede et appareil d'annulation d'interferences dans un recepteur de trafic a acces multiple

Country Status (2)

Country Link
AU (1) AU4653393A (fr)
WO (1) WO1994000918A1 (fr)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1999005833A1 (fr) * 1997-07-28 1999-02-04 Ericsson Inc. Procedes et dispositifs de demodulation conjointe de signaux de canaux adjacents dans des systemes de communication numeriques
WO1999005832A1 (fr) * 1997-07-28 1999-02-04 Ericsson Inc. Procedes et dispositif permettant d'annuler des signaux de canaux adjacents dans des systemes de communication numeriques
CN101986573A (zh) * 2010-10-25 2011-03-16 中兴通讯股份有限公司 一种双模通信系统的频谱干扰抵消装置、系统及方法
WO2018206246A1 (fr) * 2017-05-11 2018-11-15 Robert Bosch Gmbh Dispositif de traitement de signaux pour un système de communication pouvant être utilisé en particulier dans un système de batterie

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3873931A (en) * 1973-10-05 1975-03-25 Comstron Corp FM demodulator circuits

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3873931A (en) * 1973-10-05 1975-03-25 Comstron Corp FM demodulator circuits

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
M.A.BYKHOVSKIY: "SINGLE-CHANNEL RADIO INTERFERENCE COMPENSATORS IN COMMUNICATION SYSTEMS", TELECOMMUNICATIONS AND RADIO ENGINEERING, vol. 35/36, no. 11, November 1981 (1981-11-01), WASHINGTON US, pages 86 - 92 *
MILCOM 86 - PROCEEDINGS OF 1986 IEEE MILITARY COMMUNICATIONS CONFERENCE 5-9 October 1986, Monterey, California (US NEW YORK (US) *
PROCEEDINGS OF THE TWENTY-SECOND ASILOMAR CONFERENCE ON SIGNALS, SYSTEMS & COMPUTERS Pacific Grove, California (US) 31 October - 2 November 1988 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6137843A (en) * 1995-02-24 2000-10-24 Ericsson Inc. Methods and apparatus for canceling adjacent channel signals in digital communications systems
WO1999005833A1 (fr) * 1997-07-28 1999-02-04 Ericsson Inc. Procedes et dispositifs de demodulation conjointe de signaux de canaux adjacents dans des systemes de communication numeriques
WO1999005832A1 (fr) * 1997-07-28 1999-02-04 Ericsson Inc. Procedes et dispositif permettant d'annuler des signaux de canaux adjacents dans des systemes de communication numeriques
US6108517A (en) * 1997-07-28 2000-08-22 Ericsson Inc. Methods and apparatus for joint demodulation of adjacent channel signals in digital communications systems
CN101986573A (zh) * 2010-10-25 2011-03-16 中兴通讯股份有限公司 一种双模通信系统的频谱干扰抵消装置、系统及方法
WO2012055321A1 (fr) * 2010-10-25 2012-05-03 中兴通讯股份有限公司 Dispositif, système et procédé d'annulation d'interférences spectrales d'un système de communication en mode mixte
CN101986573B (zh) * 2010-10-25 2014-03-12 中兴通讯股份有限公司 一种双模通信系统的频谱干扰抵消装置、系统及方法
WO2018206246A1 (fr) * 2017-05-11 2018-11-15 Robert Bosch Gmbh Dispositif de traitement de signaux pour un système de communication pouvant être utilisé en particulier dans un système de batterie
US10873353B2 (en) 2017-05-11 2020-12-22 Robert Bosch Gmbh Signal processing device for a communication system usable in particular in a battery system

Also Published As

Publication number Publication date
AU4653393A (en) 1994-01-24

Similar Documents

Publication Publication Date Title
US5461646A (en) Synchronization apparatus for a diversity receiver
US5105435A (en) Method and apparatus for cancelling spread-spectrum noise
EP1783922B1 (fr) Système et procédé permettant d'établir une estimation précise de signaux brouilleurs reçus dans les systèmes de communication sans fil
US6498784B1 (en) Cancellation of pilot and traffic signals
CN101494472B (zh) 无线通信系统内对接收信号预测的系统和方法
US6839379B1 (en) Adaptive DS-CDMA multi-user receiver with diversity combining for interference cancellation
WO1992011722A1 (fr) Procede et appareil de suppression de bruit a spectre etale
US5479453A (en) Wireless telecommunication digital receiver
Syrjälä et al. Pilot allocation and computationally efficient non-iterative estimation of phase noise in OFDM
Min et al. Analysis and design of a frequency-hopped spread-spectrum transceiver for wireless personal communications
RU2177207C2 (ru) Приемник мобильной системы связи и способ приема в мобильной системе связи
EP0674401B1 (fr) Recepteur de signaux a etalement du spectre
EP0343189A4 (en) Tdma communications system with adaptive equalization
De Gaudenzi et al. Design of a low-complexity adaptive interference-mitigating detector for DS/SS receivers in CDMA radio networks
US5442663A (en) Residual carrier adaptive cross-polarization equalizer
Simon et al. Synchronization over rapidly time-varying multipath channel for CDMA downlink RAKE receivers in Time-Division mode
WO1994000918A1 (fr) Procede et appareil d'annulation d'interferences dans un recepteur de trafic a acces multiple
US5787128A (en) Nonlinear signal correlator and method therefor
US8000654B2 (en) Apparatus and method for canceling interference in a single antenna 1xEV-DV base station
Min Analysis and design of a frequency-hopped spread spectrum transceiver for wireless personal communications
US4380082A (en) Digital signal receiver with FM interference elimination capability
Kavehrad Performance of cross-polarized M-ary QAM signals over nondispersive fading channels
US6690713B1 (en) Tracking loop for a code division multiple access (CDMA) system
Anderson et al. Modulation considerations for a 91 Mbit/s digital radio
JP3779554B2 (ja) 周波数オフセット推定方法及び周波数オフセット推定器

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AU CA JP KR

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): AT BE CH DE DK ES FR GB GR IE IT LU MC NL PT SE

121 Ep: the epo has been informed by wipo that ep was designated in this application
DFPE Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101)
122 Ep: pct application non-entry in european phase
NENP Non-entry into the national phase

Ref country code: CA