CN100372404C - adaptive smart antenna processing method and device - Google Patents
adaptive smart antenna processing method and device Download PDFInfo
- Publication number
- CN100372404C CN100372404C CNB008196753A CN00819675A CN100372404C CN 100372404 C CN100372404 C CN 100372404C CN B008196753 A CNB008196753 A CN B008196753A CN 00819675 A CN00819675 A CN 00819675A CN 100372404 C CN100372404 C CN 100372404C
- Authority
- CN
- China
- Prior art keywords
- signal
- weight vector
- adaptive
- weight
- determining
- Prior art date
- Legal status (The legal status 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 status listed.)
- Expired - Fee Related
Links
Images
Landscapes
- Mobile Radio Communication Systems (AREA)
- Radio Transmission System (AREA)
- Variable-Direction Aerials And Aerial Arrays (AREA)
Abstract
A method and apparatus for implementing adaptive smart antenna processing in a receiving communication station including an antenna array (103) and an apparatus for adaptive smart antenna processing is described, the method and apparatus including determining (605) weight vectors for adaptive smart antenna processing. The invention can be used to advantage in low SINR environments, for example when operating in a mobile environment where the remote user is moving at high speed and the signal is fading. One aspect is hybrid weight adaptation, starting with a method with good convergence properties, e.g. a method that converges in a low SINR environment, and then switching to a method that converges quickly, e.g. when starting with a relatively high quality initial condition. To handle high mobility, weights determined from the data in a particular burst are applied to that particular burst. Such weights may not be optimal for subsequent bursts. When multiple users are in a given channel, a multi-port architecture is employed to track individual remote users.
Description
Technical Field
The present invention relates to wireless communication systems, and more particularly, to determining weights for adaptive smart antenna processing in a wireless communication receiver having an array of antenna elements and adaptive smart antenna processing means.
Technical Field
Wireless communication systems including communication stations having antenna arrays and adaptive smart antenna processing devices are well known. Such communication stations are sometimes referred to as smart antenna communication stations. When receiving signals from subscriber units, the signals received by the individual antenna array elements are combined by the adaptive smart antenna processing means to provide an estimate of the signal received from a particular subscriber unit. With smart antenna processing, including linear spatial processing, each complex-valued (i.e., including an in-phase component I and a quadrature component Q) signal received from an antenna element is weighted in amplitude and phase by a weighting factor, and then the weighted signals are summed to provide an estimate. Thus, the adaptive smart antenna processing apparatus can be described by a set of complex-valued weights, each weight corresponding to an antenna element. These complex-valued weights can thus be described by a single complex-valued vector of m elements, where m is the number of antenna elements. This can be extended to include space-time processing where the signals on the individual antenna elements are not simply weighted in amplitude and phase, but are filtered through some complex-valued filter, typically for time equalization. The individual filters may be described by complex-valued transform functions or convolution functions. Thus, the adaptive smart antenna processing of all elements can be described by the complex-valued m-vectors of the m complex-valued convolution functions.
Several methods are known for determining the weight vector of the received signal. These include methods of determining the direction of arrival of a signal transmitted by a subscriber unit and methods of using spatial characteristics of the subscriber unit, such as a spatial signature. For example, with respect to a method of utilizing the direction of arrival, see U.S. patents 5515378 and 5642353 entitled "space division multiple access wireless communication system" to Roy et al; with respect to the method of utilizing spatial signatures, reference is made to U.S. patent 5592490 entitled "efficient spectral high capacity wireless communication system" to barrat et al, and U.S. patent 5828658 entitled "efficient spectral high capacity wireless communication system with spatio-temporal processing" to Ottersten et al. So-called "blind" methods determine the weights from the signal itself, without the aid of a training signal-i.e. without determining what weights can best estimate the known symbol sequence. This method typically uses some known characteristic of the signal transmitted by the subscriber unit to determine the best weights to use by forcing estimates to have this characteristic, and is therefore referred to as a characteristic recovery method. The characteristic recovery methods can be divided into two groups. Rather than fully reconstructing the modulated received signal, for example by demodulating and then remodulating, the "partial" characteristic recovery method recovers one or more generally simple signal characteristics. The "decision directed" (DD) method constructs an accurate copy of a signal by making symbol decisions (e.g., demodulation) on the received signal.
One example of a first set of partial recovery methods is the Constant Modulus (CM) method, which is suitable for communication systems employing modulation schemes with a constant modulus, including, for example, phase Modulation (PM), frequency Modulation (FM), phase Shift Keying (PSK), and Frequency Shift Keying (FSK). See, e.g., j.r.treichler; l. larimore: "New Processing technique based on constant modulus algorithm" (IEEE Transactions on Acoustics, speech, and Signal Processing, ASSP-33, vol.2, pp.420-431, 1985, 4 months). Other partial characteristic recovery techniques include techniques that recover spectral characteristics of the signal, such as spectral auto-coherence. Spectral coherence recovery techniques exploit the known spectral coherence properties of any signal received at an antenna array. For example, in some cases, it may be assumed that the signal is cyclostationary, i.e. has a periodic self-correlation function. Other methods include recovering higher order statistics such as moments or cumulants. See, for example: "spectral self-coherence recovery" by b.agee, s.schell, w.gardner: a new method of blind adaptive signal extraction using antenna arrays "(Proceedings of the IEEE volume 78, phase 4, 1990 month 4); and us patent 5260968 to Gardner et al entitled "method and apparatus for multiplexing communication signals by blind adaptive spatial filtering"; and patent 5255210 entitled "self-coherent recovery signal extraction apparatus and method" to Gardner et al.
The decision-directed method takes advantage of the fact that the modulation scheme of the transmitted subscriber unit signal is known and determines the weights to generate a signal with the desired modulation scheme (the "reference signal"), which if transmitted by a remote user, would generate a signal at the antenna element in the array "near" the actual received signal, including making symbol decisions. For an introduction to a system employing decision-directed weight determination, see, for example: barratt et al, U.S. patent application Ser. No. 08/729390 entitled "method and apparatus for decision directed demodulation Using antenna arrays and spatial processing" (filed 10/11/1996); and 09/153110 of Petrus et al, entitled "method of generating reference signals in the presence of frequency offset in a communication station employing spatial processing" (9/15 of 1998).
It is well known that some iterative methods, including partial recovery methods such as the CM method, converge even for low signal-to-noise ratios (SNRs), low signal-to-interference-plus-noise ratios (SINRs), and high fading conditions that may be encountered in communication systems where subscriber units are moving at high speeds. Such methods are referred to herein as "iterative weight determination methods with good convergence properties". However, methods with good convergence properties can take many iterations to converge. For example, the CM method may take many iterations to converge, and may not converge fast enough in a practical system. For example, in a highly mobile system, it is preferable to use a weight vector for the current burst derived from the data of the current burst. This means that the weights are computed quickly, which may not be possible with the CM method. Decision-directed methods, on the other hand, are an example of a class of methods that converge quickly in the case of initial conditions, such as high initial signal-to-noise ratio (SNR) and signal-to-interference-plus-noise ratio (SINR), or initial weight vectors that are close enough to the correct value. The method of fast convergence in case the initial weight vector is close enough to the correct value is referred to herein as the "fast convergence iterative weight determination method". Fast convergence methods, such as the DD method, are increasingly used in smart antenna based communication stations. When this method fails in so-called low SINR or high fading conditions, the method may not converge. This problem is particularly acute in communication systems where many users are exposed to high co-channel interference, i.e. where the interference of signals in conventional channels from other subscriber units is high when receiving signals from a particular subscriber unit, such other subscriber units being from the same or a neighbouring cell in the case of a cellular system comprising several receiving communication stations, each communicating with a group of subscriber units located within its cell.
In theory, adaptive smart antenna processing allows more than one communication link to exist in a single "conventional" communication channel, provided that subscriber units sharing the same conventional channel can be spatially (or spatio-temporally) resolved. Conventional channels include frequency channels in Frequency Division Multiple Access (FDMA) systems, time slots in Time Division Multiple Access (TDMA) systems (which also typically include FDMA, so more precisely, conventional channels are time and frequency slots), and codes in Code Division Multiple Access (CDMA) systems. Thus, it can be said that a conventional channel is divided into one or more "spatial" channels, and when more than one spatial channel is present in each conventional channel, the multiplexing is referred to as Spatial Division Multiple Access (SDMA). SDMA is used herein to refer to adaptive smart antenna processing that includes one and more spatial channels per conventional channel.
Fast convergence methods, such as decision-directed methods, also fail in the presence of high co-channel interference in SDMA systems with more than one spatial channel per conventional channel.
Accordingly, there is a need in the art for an adaptive smart antenna processing method that efficiently determines adaptive smart antenna processing weights in a low signal-to-interference-plus-noise environment or in a high fading environment for both SDMA systems having one spatial channel per conventional channel and SDMA systems having multiple spatial channels per conventional channel.
Therefore, there is a need in the art for a weight determination method that performs good and fast convergence (i.e., with fewer iterations) under low SINR and high fading conditions.
Therefore, there is a need in the art for a method that combines good convergence properties with fast convergence properties.
Therefore, there is a need in the art for a "blind" method (i.e., one that does not employ training data) that combines good convergence properties (convergence at low SINR) with fast convergence.
Summary of The Invention
An object of the present invention is a weight determination method that combines the advantages of a method with good convergence properties with the advantages of a method with fast convergence.
Another object of the present invention is a "blind" weight determination method and apparatus that performs well and fast, i.e. converges in a few iterations, under low SINR and high fading conditions.
Another object is an adaptive smart antenna processing method and apparatus that efficiently determines adaptive smart antenna processing weights in low signal-to-interference-plus-noise environments or high fading environments for SDMA systems having a spatial channel for each conventional channel.
Another object is an adaptive smart antenna processing method and apparatus that efficiently determines adaptive smart antenna processing weights in low signal-to-interference-plus-noise environments or high fading environments for SDMA systems having multiple spatial channels per conventional channel.
Another object of the present invention is an adaptive smart antenna processing method and apparatus to determine the weights for adaptive smart antenna processing to be used in a current burst of data, by adapting these weights to the current burst of data by determining from the data in the current burst.
Brief description of the drawings
The present invention will be understood more fully from the detailed description of the preferred embodiments of the invention, which, however, should not be taken to limit the invention to any particular embodiment, but are for explanation and better understanding only. These embodiments are again illustrated by means of the following figures:
fig. 1 is a functional block diagram of a multi-antenna transceiver system in which a receive right determining apparatus according to aspects of the present invention may be included;
fig. 2 is a more detailed block diagram of a transceiver including a signal processor that, when executing a set of instructions, implements a receive right determination apparatus in accordance with aspects of the present invention;
FIG. 3 is a flow chart of one embodiment of a rights determination method of the present invention;
FIG. 4 shows a block diagram of a tracking reference signal generator and demodulator employed in a preferred embodiment of the present invention;
FIGS. 5A, 5B, and 5C illustrate the performance of a constant modulus method, a decision directed method, and a hybrid method according to aspects of the present invention, respectively;
FIG. 6 is a block diagram of a multi-port weight determination apparatus and spatial processor in accordance with a preferred embodiment of the present invention;
FIG. 7 is a flow chart of a preferred embodiment of the multi-user weight determination method of the present invention;
FIG. 8 illustrates the effect of timing offset on CM method performance; and
fig. 9 illustrates a block diagram of an apparatus implementing one aspect of the disclosure.
Detailed description of the preferred embodiments
Architecture of a base station
The method and apparatus of the preferred embodiment are implemented in a communication receiver, and in particular, in a PHS-based antenna array communication station (transceiver), as shown in fig. 1, where there are m antenna elements in the antenna array. In this particular embodiment, m =4. While systems similar to that shown in FIG. 1 may be prior art, systems such as the system of FIG. 1 having elements that implement aspects of the present invention either programmatically or hardwired are not prior art. Furthermore, the present invention is in no way limited to use with PHS air interfaces or TDMA systems, but rather any communication receiver that includes an adaptive smart antenna processing device. In fig. 1, a transmit/receive ("TR") switch 107 is connected between the m-antenna array 103 and transmit electronics 113 (including one or more transmit signal processors 119 and m transmitters 120) and receive electronics 121 (including m receivers 122 and one or more receive signal processors 123) to selectively connect one or more elements of the antenna array 103 to the transmit electronics 113 when in a transmit mode and to selectively connect one or more elements of the antenna array 103 to the receive electronics 121 when in a receive mode. Two possible implementations of the switch 107 are a frequency diplexer in a Frequency Division Duplex (FDD) system and a time switch in a Time Division Duplex (TDD) system. The PHS preferred embodiment of the present invention employs TDD. The transmitter 120 and receiver 122 may be implemented using analog electronics, digital electronics, or a combination of both. The receiver 122 of the preferred embodiment generates digitized signals that are fed to one or more signal processors 123. The signal processors 119 and 123 may be static (always the same), dynamic (varying according to a desired directivity) or intelligent (varying according to the received signal), and in a preferred embodiment are adaptive. The signal processors 119 and 123 may be the same or multiple DSP devices programmed differently for reception and transmission, or different DSP devices, or different devices for some functions and the same device for other functions.
It should be noted that although fig. 1 shows a transceiver in which the same antenna elements are used for receiving and transmitting, it is clear that separate antennas may be used for receiving and transmitting, and that either only receiving or only transmitting or both receiving and transmitting may involve adaptive smart antenna processing.
For example, version 2 of the radio industry and commerce association (ARIB, japan) major standard, the Personal Handyphone System (PHS) described in the RCR STD-28, and the modification described in the PHS memorandum group understanding technical standard (PHS MoU-see http:// www.phsmou.or. Jp), are 8-slot Time Division Multiple Access (TDMA) systems with actual Time Division Duplex (TDD). Thus, 8 slots are divided into 4 Transmit (TX) slots and 4 Receive (RX) slots. This means that for any one particular channel, the receive frequency is the same as the transmit frequency. This also implies reciprocity, i.e. the propagation paths used for the downlink (from the base station to the user's remote terminal) and the uplink (from the user's remote terminal to the base station) are identical, assuming minimal movement of the subscriber unit between the receive and transmit time slots. The frequency band of the PHS system used in the preferred embodiment is 1895-1918.1MHz. Each of the 8 slots is 625 microseconds long. The PHS system has a dedicated frequency and time slot for the control channel on which the call initialization is performed. Once the link is established, the call is transferred to the traffic channel for conventional communication. Communication takes place in any channel at a rate of 32 kilobits per second (kbps), referred to as full rate. Sub-full rate communications are also possible and it will be clear to those skilled in the art how to modify the embodiments described herein to incorporate sub-full rate communications.
In PHS as employed in the preferred embodiment, a "burst" is defined as an RF signal of limited duration transmitted or received over the air during a single time slot. A "group" is defined as a group of 4 TX slots and 4 RX slots. One group always starts from the first TX slot and has a duration of 8 × 0.625=5 microseconds.
The PHS system employs pi/4 differential quadrature phase shift keying (/ 4 DQPSK) modulation of the baseband signal. The baud rate is 192 kbauts. I.e., 192000 symbols per second.
Fig. 2 shows a block diagram of a more detailed, but still simplified PHS base station including adaptive smart antenna processing and upon which an embodiment of the present invention is implemented. Further, while systems having a structure similar to that shown in FIG. 2 may be prior art, systems such as the system of FIG. 2 having elements programmed or hardwired to implement aspects of the present invention are not prior art. In fig. 2, a plurality of m antennas 103 is employed, where m =4. More or fewer antenna elements may be used. The output of the antenna is connected to a duplexer switch 107, which in this TDD system is a time switch. When received, the antenna output is coupled to the receiver 205 through the switch 107, mixed down from the carrier frequency (about 1.9 GHz) to an intermediate frequency ("IF") in an analog manner by the RF receiver module 205. This signal is then digitized (sampled) by an analog-to-digital converter ("ADC") 209. Digital down conversion is then performed by digital down converter 213 to generate a four times oversampled complex-valued (in-phase I and quadrature Q) sampled signal. Thus, elements 205, 209 and 213 correspond to receiver 122 of fig. 1. For each of the m receive time slots, the m down-converted outputs from the m antennas are fed to a Digital Signal Processor (DSP) device 217 (hereinafter referred to as a "time slot processor") for further processing. In the preferred embodiment, a commercially available DSP device is used as the slot handler, one DSP for each receive slot.
The timeslot processors 217 perform a variety of functions including: received signal power monitoring, frequency offset estimation/correction and time offset estimation/correction, smart antenna processing including determining weights for individual antenna elements using a method according to an aspect of the present invention to determine a signal from a particular remote user, and demodulation of the determined signal.
The output of the slot processor 217 is a demodulated data burst corresponding to each of m (= 4) receive slots. This data is sent to the main DSP processor 231 whose primary function is to control all elements of the system and to cooperate with the higher layer processing of what signals are required for communication in all of the different control and traffic communication channels defined in the PHS communication protocol. In the preferred embodiment, the host DSP 231 is also a commercially available DSP device. In addition, the slot processor transmits the determined reception right to the host DSP 231.
The transmit controller/modulator 237 receives transmit data from the host DSP 231. The transmit controller uses this data to generate an analog IF output to be sent to the RF Transmit (TX) module 245. Certain operations performed by the transmit controller/modulator 237 include converting data bits to complex-valued pi/4 DQPSK modulated signals, up-converting to IF frequencies, weighting by complex-valued transmit weights obtained by the host DSP 231, and converting the signals to analog transmit waveforms to be transmitted to the transmit module 245 using a digital-to-analog converter ("DAC"). The transmit module 245 up-converts the signals to a transmit frequency and amplifies the signals. The amplified transmit signal output is coupled to the m antennas 103 through a duplexer/timing switch 107.
Symbol
The following notation is used. Suppose there are m antenna elements (m =4 in the preferred embodiment) and z 1 (t)、z 2 (t)、...z m (t) is the complex response (i.e. with in-phase I and quadrature Q components) of the first, second,. Eta.m antenna element after down-conversion, i.e. in baseband and after sampling (in the preferred embodiment four times over-sampling), respectively. In the above notation, but not essential to the invention, t is discrete. These m-time sample quantities may be represented by a single m-vector z (t), where z (t) in row i is z i (t) of (d). For each burst, a limited number, say N, of samples are collected, such that z 1 (t)、z 2 (t)、...z m (t) may each be represented as an N row vector and Z (t) may be represented by an m N matrix Z. In the following detailed description, details incorporating a limited number of samples are omitted, and it will be clear to one skilled in the art how these details are included.
Assuming that the plurality of signals is from a number, say N s A remote user transmits to a base station. In particular, it is assumed that the subscriber unit concerned transmits a signal s (t). Adaptive smart antenna processing includes processing a received signal z 1 (t)、z 2 (t)、...z m The values of I and Q of (t) are specifically combined to extract an estimate of the transmitted signal s (t). These weights may be represented by a receive weight vector corresponding to a particular subscriber unit, represented as having an ith element w ri Of complex-valued weight vectors w r . Then the message is sentThe number estimates are:
wherein w ri ' is w ri Is complex conjugated, and w r H Is a receiving weight vector w r Hermitian transpose (i.e., transpose and complex conjugate). In embodiments that include spatio-temporal processing, each of the weight vectors is receivedThe elements are functions of time, so the weight vector can be expressed as having the ith element w ri W of (t) r (t) of (d). The estimate of the signal can be expressed as:
wherein the operator " * "is a convolution operation. Spatio-temporal processing, for example combining temporal equalization with spatial processing, is particularly useful for wideband signals. The estimation of the signal using space-time processing can be equivalently performed in the frequency domain (fourier transform).
z i (t) and w ri (t) frequency domain representations are respectively
Z i (k) And W i (k) Where k is the value of the discrete frequency,
with space-time processing, the convolution operation of equation (2) is typically finite and performed on sampled data, which is equivalent to combining spatial processing with temporal equalization using a time-domain equalizer with a finite number of equalizer taps. I.e. each w ri (t) has a finite number of values of t, corresponding to each W in the frequency domain i (k) With a limited number of k values. If the convolution function w ri If the length of (t) is n, the complex m-weight vector w is not determined r Instead, a complex-valued mxn matrix W is determined r Its column is w r N values of (t).
In the rest of the description, whenever reference is made to a complex-valued receive weight vector w r Or elements thereof, it is to be understood that this may be used in spatial processing or broadly in conjunction with spatio-temporal processing as described above to determine the weight matrix W r . Thus, both spatial processing and spatio-temporal processing are referred to herein as adaptive smart antenna processing.
Determining weights of space
A "blind" method of determining weights for adaptive smart antenna processing is to not require reconstruction of the training data. The method of the present invention, like most blind methods, takes advantage of some knowledge of the format of the originating signal and limits the output signal to have one or more known input signal characteristics. The characteristic may be an amplitude characteristic or some statistical characteristic, such as entropy or cyclostationarity, correct modulation scheme, or reconstruct an exact replica. Such a method is sometimes referred to as "property recovery".
Method with good convergence properties
The method having a good convergence characteristic includes a partial characteristic recovery method: some methods of reproducing one or more characteristics by determining a bitstream and reconstructing a signal without attempting to reproduce an exact copy. Such methods include methods that preserve the amplitude (mode), entropy, and spectral coherence (e.g., cyclic stability) of the signal.
The Constant Modulus (CM) method is a very simple and efficient technique that is applicable to signals modulated by schemes that result in constant amplitude signals. These include all forms of phase and frequency modulation, including differential phase shift keying modulation of the PHS system employed in the preferred embodiment. The CM method is also applicable to non-constant modulus signals, as described below. The CM method determines the weights that satisfy the following conditions: a) Restoring the constant amplitude (constant modulus) characteristic of the signal, and b) generating the signal, if transmitted by a remote user, should be generated at the antenna elements of the array "close" to the actual received signal. Amplitude variations may be introduced due to interference, fading, and timing offsets in those modulation schemes where the constant modulus characteristic depends on accurate timing offset correction, including, for example, the DQPSK modulation scheme of the preferred embodiment where the constant modulus characteristic only remains at the baud point. In the presence of co-channel interference, the CM method tends to pick up the strongest signal, whether it be the desired signal or co-channel interference. Even if the strength of the desired signal is only 0.5db greater than the strength of any interfering signals, the cm method will still correctly pick up the strongest, i.e., desired, signal. That is, the CM method has very good convergence characteristics.
There are many variations of the CM method. They generally minimize a cost function of general form:
where E (.) denotes a statistical expectation operation, and p and q are positive integers, typically 1 or 2. It will be apparent to those skilled in the art that in practice the statistical operation is replaced by some form of sample averaging or accumulation (e.g. by summing a set of samples which in the preferred embodiment are a subset of all samples in a burst). Again, it is clear that more terms can be added to the cost function of equation (4) without going beyond the scope of the present invention. For example, a term may be added to limit the size of the weight vector. See the above-cited U.S. patent application 08/729390 for an example of a cost function (non-CM cost function) with such an added term. Signal S ref (t) is the normalized replica signal (referred to as the "reference signal") employed in the cost function. I.e. forThe reference signal from which the weights are determined is a weighted sum of the receive antenna signals which are then normalized. The determination of the weights is to determine a set of weights that minimizes the cost function in equation (4).
The CM method is also applicable to non-constant modulus signals. See, for example, j.lundell and b. Widrow: "application of constant modulus adaptive beamforming devices to constant and non-constant modulus signals" [ Proceedings,1988 agilomar Conference on signals, systems and Computers (ACSSC-1988), pp 432-436, 1988 ]. Lundell and Widrow use cost functions such as equation (4) for p = q =2 (this is called the 2-2CM method), demonstrating that any constant and non-constant modulus signal can be recovered using this 2-2CM method as long as the ratio of the fourth moment to the square of the second moment (this ratio is called the kurtosis) is below 2. For example, it is well known that M-Quadrature amplitude modulated signals (M-QAM) have peak states of about 1.4-1.2/(M-1), and thus the peak state of any QAM signal is always less than 1.4. The CM method is applicable to such signals.
The particular method with good convergence properties, at least one iteration of the CM method, is used in the preferred embodiment, and the implementation of the CM method uses values for p being 1 and q being 2 in equation (4). When the invention is applied to non-constant modulus signals and the CM method is employed, other values of p and q may also be employed, for example p = q =2. The best implementation may also be a block-based approach. I.e. a block of signals received by the antennas is weighted and the weights are determined using this block. A block is a subset of the samples in a burst. Specifically, it is preferable to use 75 samples of 120 PHS burst symbols, where the 75 symbols are in the payload in the middle of the PHS burst. The use of data in the payload advantageously ensures that the data used for the calculation of the rights for any one remote user is different from the data used by another user unit. There are a maximum of 88 such payload samples in a PHS burst.
When p =1 and q =2, this method is called least squares constant modulus method, comprising the following steps:
1. the weight vector is initialized. For example, using w r，initial ＝[100...0]Where x' represents the transpose of x. In an improved embodiment, the R corresponding to the maximum singular value of Z is adopted zz ＝ZZ H The largest eigenvector of (a). In yet another embodiment, a weight vector derived from a previous burst is employed;
2. for the samples involved, a replica signal and normalization are performed:
3. computing a weight vector w using a least squares method r . That is to say that the first and second electrodes,
where N is the number of samples used in the calculation. The solution of equation (6) is:
wherein R is ZZ ＝ZZ H ，
And N is the number of samples used; and
4. repeating step 2 and step 3 until convergence is reached.
It should be noted that in the calculation of step 3, in practice, the overall scaling factor is not important. The proprietary scaling factors are preferably applied in combination as gains to the system.
It should be noted that the CM method can be extended to spatio-temporal processing. One well-known method employs a 2-2CM method, indicating that the CM method used for spatio-temporal weight determination (i.e., for weight matrix determination) must converge under some assumption that the actual situation is usually met. See c.b. Papadias and a.paulraj for spatio-temporal constant modulus algorithms for SDMA systems (Proceedings, IEEE46th vehicular technology conference, pages 86-90, 1996), but this method is not based on block data. However, the spatial weight determination method can be easily modified for space-time processing according to the weight matrix by restating the problem according to matrices and vectors of different sizes. For the entire description, it is assumed that m is the number of antenna elements and N is the number of samples. Let n be the number of time equalizer taps per antenna element. Can be used forA vector of N samples per row of an (m × N) received signal matrix Z is rewritten to an offset version of a first row of N rows to generate a received signal matrix Z of size (mn × N), and an estimated received signal row vector of N samples is generated when a Hermite transpose of a weight vector of size (mn × 1) is left-multiplied. Thus, the spatio-temporal problem has been reformulated as a weight vector determination problem. For the CM method, in equation (7), for example, the weight vector is a "long" weight vector of size (mn × 1), R zz Is a matrix of size (mn × mn), and r ZS Is a long vector of size (mn × 1). Rearranging the terms provides the required (m × n) weight matrix.
Since in the preferred embodiment the sampled data needs to be substantially baud (on-baud) in order to maintain the CM characteristics, the timing offset estimation and correction is performed while step 2 is performed, which in this case may include sampling and interpolation in time, since the samples of the signals received from the individual antenna elements of the antenna array 103 are oversampled and may include some timing offset. Thus, the variable t in equations (5) and (7) represents the time of the sample approximately over baud. Obviously and as described in the above-cited us patent application 09/153110, the timing offset estimation/correction (which may include sampling/interpolation) may be performed on the m signals prior to the signal replication operation, or may also be performed after the signal replication operation.
Simulations have been performed to determine the level of accuracy of the timing offset/wave characteristic estimation. Fig. 8 shows the results. In this simulation, the output SINR obtained with the CM weights is calculated without sampling the signal exactly on the wave characteristics, but offset from the ideal baud point by some timing offset, which varies from-1/2 to +1/2 baud by a step size of 1/8 of the baud. The results show that the output SINR only decreases by 0.6dB even for signal offsets at ± 1/8 baud. Although this number is for the test example shown, the conclusion is that the accuracy of the timing offset correction of the CM method need not be very high. Thus, a simple method can be used to perform offset correction/sampling/interpolation to generate samples that are approximately baud aligned.
It should also be noted that timing offset correction (including sampling/interpolation) of an oversampled signal may not be necessary for all modulation schemes with constant modulus characteristics. For example, the "pan-european digital cordless telecommunications" (DECT) standard and the "global system for mobile communications" (GSM) standard employ Gaussian Minimum Shift Keying (GMSK) signals that always have a constant modulus, so timing offset correction is not required for the CM method in those instances.
The least squares CM weight determination method is quite simple to implement. Because no demodulation is performed, no frequency offset estimation and correction and demodulation is required. In one embodiment of the present invention, although not required, frequency offset correction is performed when the CM method is implemented. An additional feature of simple property recovery methods such as the constant modulus method is that convergence occurs even if the value of the signal-to-interference-plus-noise ratio (SINR) is low.
The main drawback of methods with good convergence properties, such as simple property recovery methods, is that they may take many iterations to converge. In a typical system, the processing power, e.g., of a DSP, is extremely limited, so that, for example, in order to utilize the weight vectors within the current burst, convergence may not occur with the CM method for a certain amount of time.
Fast convergence method
Unlike partial feature recovery methods, fast convergence methods, such as decision-directed methods, converge very fast. With the decision directed approach, the recovered characteristic is a complete copy of the original signal with the correct modulation scheme. That is, a signal replication operation, estimating a received signal as in equation (1), demodulating the signal and constructing a reference signal with a correct bit stream. To do so, the reference signal is constructed with any frequency and timing offsets that need to be corrected. The correct weights generate a reference signal that approximates the transmitted signal. The scheme may involve one or more iterations to obtain the "best" weights. Although the timing offset correction (including any samples) and the frequency offset correction are shown below as occurring after the signal duplication operation, it will be apparent that one or more of these may occur prior to the signal duplication operation. See above-referenced commonly owned U.S. patent applications 08/729390 and 09/153110 for a detailed description of examples of these operations performed before and after signal replication, and decision-directed methods. When the least squares criterion is employed, the general method comprises the steps of:
1. the weight vector is initialized. For example, using w r，initial ＝[100...0]Where x' represents the transpose of x. In a modified embodiment, R corresponding to the largest singular value is used ZZ ＝ZZ H The singular vector of (a). In yet another embodiment, a weight vector derived from a previous burst is employed. As described below, one aspect of the present invention includes employing a decision directed approach after employing a partial characterization approach. In this case, the resulting weight vector is used (i.e., the partial property recovery method is used) when implemented in any of the embodiments of the present invention;
2. performing signal replication
If the sample is initially oversampled, then sampling/interpolating (in another arrangement, sampling/interpolating may be performed before the replica signal operation);
3. estimating timing and frequency offsets to produce a signal with correct timing and frequency offsets;
4. by making symbol decisions (i.e. demodulation), so that S ref (t) has the correct bit stream and the same modulation scheme, and has the same signal as that transmitted from a particular user to the receiverTiming and frequency offset to determine the reference signal S ref (t)；
5. By exceeding w r The weight vector is calculated by least squares minimization. That is to say that the first and second electrodes,
the solution of the method is that,
wherein R is ZZ ＝ZZ H And
(ii) a And
6. steps 2, 3, 4 and 5 are repeated until convergence is reached.
It should be noted that steps 2, 3 and 4 require the signal to be corrected for frequency and timing offsets in order to make the correct demodulation decision at step 4, while step 5 typically requires the correct frequency and timing offsets to be reintroduced so that the reference signal and replica signal in the cost function have the same timing and frequency offsets. It is further noted that the minimization of equation (9) may also include other terms, such as a weighted norm term for the weight vector, to place constraints on the norm of the weight vector, as discussed above for the CM example and in the U.S. application 08/729390 cited above. See also the above-cited commonly owned U.S. patent applications 08/729390 and 09/153110 for a detailed description of how to determine the reference signal (step (4)). This will be explained below with the aid of fig. 4.
It should be noted that the decision-directed approach can be readily extended to determine the weight matrix for spatio-temporal processing, e.g., rearranging terms for the CM approach as described above, and by other approaches well known to those skilled in the art. Therefore, the present invention also includes a method for determining a weight vector and a weight matrix for a timing null process.
Thus, the decision-directed approach reproduces an exact replica of the signal that is supposed to be transmitted to the receiver, while the partial characteristic recovery approach reproduces one or more simple characteristics, such as the correct amplitude. Decision-directed systems perform very well and converge with very few iterations in environments with fairly high SINR. However, these methods are sensitive to initial conditions and cannot even converge when the initial SINR is low. This is common in high mobility cellular systems and other systems that exhibit fading.
It should be noted that iteration of the CM method is generally computationally less costly than iteration of the decision-directed method, since no frequency offset correction or demodulation is required.
The best method comprises the following steps: single user
One aspect of the present invention is a weight determination method comprising N performing an iterative weight determination method (such as a partial characteristic recovery method, preferably a CM method) with good convergence characteristics 1 A second number N of iterations combined with a fast convergence method and then performing the fast convergence method (such as a decision-directed method) 2 In order to obtain the advantages of good convergence characteristics and fast convergence. N is a radical of hydrogen 1 N of one CM iteration decision-directed method 2 The starting condition of the iteration is in the most likely range of fast convergence of the decision-directed method. The preferred embodiment employs one iteration (N) of the decision-directed approach 2 = 1), while another implementation uses two iterations (N) 2 = 2). This method can be re-described as performing iterations of the iterative weight determination method with good convergence properties until the switching criterion is met, and then performing several iterations of the fast convergence method starting with weights obtained with the method with good convergence properties. In one example, the switching criterion is a well-defined number of iterations N 1 . In another preferred embodiment, N 1 Not explicitly specified. In contrast, the switching criterion is a SINR threshold for the replica signal, and when the SINR estimate equals or exceeds the threshold, switching is made to the decision directed approach. Thus, a sufficient number N is utilized 1 To obtain sufficient information to ensure that only N is used 2 This further iteration causes the decision-directed method to converge on the SINR.
Many methods of determining the SINR estimate may be employed. In the preferred embodiment, the method used is as described in Yun, U.S. patent application Ser. No. 09/020049 entitled "Power control with Signal quality estimation for Smart antenna communication systems" (filed 2/6/1998). The implementation of the signal quality estimation method will now be described.
The number of samples of the burst to be used for estimation is denoted by N. The sampled modulus information is first extracted by summing the squares of the in-phase and quadrature signals (the real and imaginary parts of the signal s (t)). The average power and the average squared power are then determined by averaging those number of samples for which operation is desired.
It should be noted that once the instantaneous power R is determined 2 (t)＝I 2 (t)+Q 2 (t) determining the squared power R 4 (t)＝[R 2 (t)] 2 Only a single additional multiplication is required for each sample and the estimated signal-to-interference-plus-noise ratio is determined, preferably using at most one square root operation, using
Ratio of
And quantity A are sometimes referred to as kurtosis. This preferred embodiment of the signal quality estimation is not sensitive to frequency offset, so it is a particularly attractive approach to use with CM methods that are also not sensitive to frequency offset.
In another implementation of the invention, other methods of determining the quality of the post-copy operation signal may also be employed.
A method of determining rights to a single user is illustrated in the flow chart of fig. 3. An initial weight vector is formed at 303. It may be [100.. 0]' or in a modified embodiment, the R corresponding to the largest singular value is taken ZZ ＝ZZ H The singular vector of (a). In yet another embodiment, the weight vector in the previous burst is used. The copy operation 305 is now performed according to equation (1), but in the preferred embodiment, only the middle part of the burst is used, preferably only the 75 symbols (300 samples) in the payload part in the middle of the burst. The output is corrected for timing offsets at 307. Any timing offset correction method may be employed. As discussed above, the timing offset correction need not be very accurate. The most preferred method is as described in U.S. patent application Ser. No. 09/153110, referenced above. The copy operation and the timing offset correction operation may be combined. Any necessary sampling and interpolation is inherent in the timing offset correction operation, but is not explicitly shown in fig. 3, so after step 307, the data comprises 75 complex-valued (I and Q) samples of the wave characteristics located at about 75 symbols from the middle of the current burst. The SINR of the replica signal is preferably estimated 309 using kurtosis as described above. In step 311, it is determined whether the SINR exceeds a thresholdThe value SNR. If not, then in step 313, an iteration of the constant modulus method is performed using the least squares cost function criterion described in equations (6) and (7) above. The method then returns to the copy operation of step 305 for another iteration. On the other hand, if it is determined in step 311 that the SINR threshold has been exceeded, in steps 315 and 317, N of the decision directed method is performed 2 An iteration including a frequency offset correction at 315. Any frequency offset correction method may be used, the most preferred method being the method described in the above-referenced us patent application 09/153110. Likewise, any method may be employed for decision directed adaptation, including generating reference signals, with the preferred embodiments employing the aboveThe method described in U.S. patent application Ser. No. 09/153110, incorporated by reference above. When weights are determined, in the preferred embodiment, only a subset of the samples in each burst is used. Thus, the finally determined weight vector is now applied to the entire burst in a copy operation and demodulation step 318. In this embodiment, step 318 includes timing and frequency offset determination and correction and demodulation, preferably using the architecture described in FIG. 4. The output of the decision directed adaptation is signal 319.
The preferred embodiment of reference signal generation as part of the decision directed adaptation of step 317 (using a portion of the burst data) and for demodulating all bursts at step 318 preferably employs a reference signal generation architecture and method that includes a tracking mechanism (preferably sample to sample) that forms the phase of the reference signal at the sample point by relaxing the phase of the signal that ideally leads the previous reference signal sample to the phase of the replica signal at the same sample point, the replica signal being formed from the received antenna signal. Constructing a reference signal at each sample point by: an ideal signal sample is constructed from the replica signal at the same sample point, the ideal signal sample having a phase determined from the replica signal at the same sample point, and the phase of the ideal signal sample at the initial symbol point is set as the initial ideal signal phase, and then the phase of the ideal signal sample is relaxed to the phase of the replica signal sample to obtain the phase of the reference signal. The phase of the ideal signal is determined based on the phase of the reference signal at the previous sample point for which the phase was determined and a decision based on the replica signal. In one implementation, the reference signal is determined in the forward time direction, and in another implementation, the reference signal samples are determined in the reverse time direction. In one arrangement, the phase of the ideal signal sample is relaxed to the phase b of the replica signal N (n) corresponds toTo which a filtered version of the difference between the phase of the replica signal and the phase of the ideal signal is added. In another arrangement, pairs of steps for relaxing the phase of the ideal signal sample to the phase of the replica signalThe reference signal samples should be formed by adding a filtered version of the difference between the replica signal and the ideal signal to the ideal signal samples.
This is now explained in more detail by means of fig. 4, taking the/4 DQPSKPHS signal as an example. It will be apparent to those skilled in the art that modifications may be made to achieve other modulation schemes. The phase detector unit 403 detects the phase difference 405 between the replica signal 401 (corrected for timing and frequency offset) and the previous reference signal 417. The phase difference signal 405 is fed to a slicer 407 to generate a decision phase difference 419. The correct phase difference for/4 DQPSK is (2 i-1)/4,i =1, 2, 3 or 4, and is the phase difference between the previous reference signal sample and the ideal signal. It is subtracted from the actual phase difference 405 to generate an error signal 411 at block 409. This error signal is filtered at filter 413 to produce filtered error signal 415. Which is a filtered error signal used to adjust the phase difference 419 closer to the actual phase difference 405. The corrected phase difference 421 is then used in a frequency synthesizer/phase accumulator 423 to generate a reference signal 429. Which is the previous sample value 417 of the reference signal 429 used by the phase detector 403, a unit time delay 425 is indicated between these signals. Symbol (N) of signal 430 2 Signal 319 after one iteration) is determined by block 427. Mathematically, if b R (t) represents the complex sample of the reference signal at the wave characteristic t and represents the phase, to the input b of the phase accumulator 423 R (t)-b R (t-1) is:
filter{d ideal (n)-decide{d ideal (n)}}+decide{d ideal (n) }, wherein decide { d } ideal (n) } is the output of the slicer 407, and is equal to (2 i-1)/4,i =1, 2, 3, or 4 for/4 DQPSK. Here, the "ideal" complex-valued sample points b ideal (t) is defined as:
b ideal (0)＝b R (0)＝b(0)，
where b (t) is the sample of the input signal 401 and d ideal (t) is the current input sample and the previous reference signalPhase difference between samples:
wherein denotes a complex conjugate. The "ideal" signal is a reference with a phase lead by an ideal amountSignal, said quantity depending on d ideal (t) the decision made. That is to say that the position of the first electrode,
b ideal (t)＝b R (t-1) + (2 i-1)/4,i =1, 2, 3 or 4. To obtain this reference signal, we will now turn b ideal (t) relaxes to b (t), in particular by the pair of quantities [ b (t) -b ] ideal (t)]I.e. b (t) and b ideal (t) filtering the phase error between (t) and adding the filtered amount to b ideal The phase of (t). Another embodiment is directed to the amounts (b (t) -b ideal (t)) instead of the phase difference. The filter is preferably a proportionality constant. Higher order filters may also be employed. Mathematically, the first and second sets of instructions, in one embodiment,
b R (t)＝b ideal (t)+filter{b(t)-b ideal (t)}，
in yet another embodiment, the architecture of FIG. 4 may be slightly modified to employ
b R (t)＝b ideal (t)+filter{b(t)-b ideal (t)}
In the preferred embodiment, the method of the flow chart of FIG. 3 includes a tracking reference signal generator, which is implemented in the form of an instruction set for the timeslot processors 217 as a signal processor (DSP) device.
The simulation of the method of FIG. 3 is performed for the system, but with an initial weight vector, R ZZ ＝ZZ H Corresponds to the largest eigenvalue. The simulation is performed for the PHS base station of fig. 2 with four antenna elements. The input signal to each antenna has a signal-to-noise ratio (SNR) of 11.9dB. The input carrier-to-interference ratio (CIR) is 1.1dB, corresponding to an initial replica signal SINR of 0.8 dB. Although a conventional PHS burst has 120 symbols, the PHS burst is not limited to a single symbolThe weight calculation is performed using only the central 75 symbols. All computations are performed offline using a MATLAB environment (Mathworks inc., natick, MA). Fig. 5A, 5B and 5C show the results comparing the convergence characteristics of the decision-directed method, the least squares CM method and the combining method of the present invention under low SINR conditions (in this case, the post-copy SINR after the first copy operation with the initial weight vector is about 0.8 dB). The output SINR (dB) after each iteration is measured and plotted using SINR estimation. The first SINR value shown is when an initial weight vector (R) is utilized ZZ The first singular vector of (c) after the replica operation. This is the same for all three methods. As shown in fig. 5A, the decision-directed method does not converge even after 10 iterations. Fig. 5B shows that the CM method converges slowly and the output (estimated) SINR increases continuously as the iteration progresses. The best SINR is 18dB and is,the CM method requires more than 10 iterations to converge to this optimum. Fig. 5C shows that the method of the present invention operates with a switching output SINR threshold of 7.5 dB. Note that this result begins to be identical to that of fig. 5B, but diverges after switching to the decision-directed approach (the result of fig. 5B is shown in dashed lines after switching). After the decision-directed method starts, the method reaches the optimal SINR after only 2 iterations of the decision-directed method, even after only one iteration of the decision-directed method is very close to the optimal value. In summary, the method of the flowchart of fig. 3 converges within 5 iterations, whereas more than 10 iterations are required if only CM is employed.
Multi-port architecture
When there are more than one, say N, in and out of a cell s When the subscriber units are in the same conventional channel (i.e., co-channel users), the preferred embodiment of the present invention employs a multi-port architecture, with each "port" forming a replica signal and tracking N s A single one of the subscriber units, so that the subscriber unit of any one port becomes the remaining N s Co-channel interference for each subscriber unit and its corresponding port. So only the tracking toolThere are co-channel users that receive signal components above a certain noise floor at the antenna element. The number of such users can be estimated. Given any one burst (matrix Z), R can be checked ZZ ＝ZZ H And may perform a first order estimation. Any order estimation method may also be used. For example, the Rissanen Minimum Description Length (MDL) criterion or the Akaike information theory criterion is well known. Overview of the technology for determining the number of active co-channel users, see section 3.8 of "estimation of direction of arrival with antenna array" by RiasMuhamed and T.S. Rappaport [ technical report MPRG-TR-96-03, mobile and Portable radio research group, bradley department of electronics, virginiaPolytechnology institute,1 month 1996]There is also RiasMuhamed's Master paper "estimation of direction of arrival using antenna arrays" [ Bradley department of Electrical engineering, virginia Polytechnology institute and Stateuniversity, blacksburg, VA24061, USA). The preferred embodiment uses a minimum description length criterion.
While the preferred embodiment involves estimating and then tracking all "active" co-channel users, in another embodiment, a "good" wireless design environment is assumed,that is, given that co-channel users that are not communicating with the same base station are far away, only valid co-channel users are those that share the same conventional channel and communicate with the base station. I.e. subscriber units of different spatial channels in the conventional channel. In this case, N s Are known.
For example, consider when there are two known (by estimation or by existing information) co-channel users (i.e., N) in this environment s = 2). When tracking one of the two subscriber units, the other subscriber unit is an interferer. Thus, in this architecture, both the desired signal and the effective interference communicating in the same conventional communication channel are tracked simultaneously. In fading environments, such as those encountered when subscriber units are moving at fast speeds, the carrier-to-interference ratio (CIR) may be low, while the instantaneous CIR may beCan fluctuate over a wide range. Thus, at any point in time, in any given burst, the desired signal may be weaker than any of its interfering signals, and the port of the desired signal may be interference locked. That is, tracking of the interferer may begin instead of the desired remote user.
Fig. 6 shows a block diagram of a multi-port adaptive smart antenna processing apparatus according to a preferred embodiment. In each port, the initial utilization corresponds to the 1 st,. N s Initial weight vectors 631-i, i =1 s The oversampled outputs 605 from the receivers 122 of the antenna elements 103 are combined in a signal copy operation 607, these initial weights being provided by a weight initialization operator 621. The resulting replica signal is time offset corrected by a time offset corrector unit 609 which also samples/interpolates to generate a set of approximately baud-aligned samples (for iterations of the CM method) or substantially baud-aligned samples (for iterations of the decision-directed method). The baud aligned samples are fed to SINR estimator 613 and the output is fed to weight calculator and demodulator 615 which uses the baud aligned samples and/or antenna signal 605 to determine a reference signal and a set of weights in accordance with the inventive method described herein. Because at least one iteration of the decision directed method is used in the weight calculator and demodulator 615, the output is a demodulated signal 617. Thus, N is determined s N of subscriber units s And a demodulation signal. Having multiple ports allows simultaneous tracking of any desired subscriber unit transmitted signal and any co-channel interferer. The adaptation method described further below enables handover between any desired user and the interferer in a fading environment. So utilize N s A port, N s The individual users being tracked simultaneously, if anySignals hop from port to port, which may occur in a fading environment, the output of the ports is classified in user classifier 623 to separate the desired user from any interfering party, outputting N correctly s A demodulated signal 625。
Other multi-port architectures are known, but are not used with the adaptive approach described herein. See, e.g., b.g. agene, "blind separation and acquisition of communication signals using multi-target constant modulus beamforming device" [1989 IEEE militariy Communications Conference ("MILCOM 89"), vol 2, 340-346, new york: IEEE,1989 ], multiport architecture for constant modulus approach. The method of Agee differs from the methods described herein in many ways, including, for example, 1) in the weight initialization method; 2) What calculations are performed in the various ports. The method of Agee orthogonalizes the ownership vectors collectively in each stage of iteration, which is computationally expensive; the preferred embodiment of the present invention allows each port to be independently adapted after the ports are co-initialized; 3) The method of determining the weight vector is different. It should be noted that in alternative embodiments, the weight vectors for the various ports are made orthogonal, as future computational power is expected to become more readily available.
The optimal weight determination method comprises the following steps: multiple users
Fig. 7 illustrates, by way of a flow chart, a preferred method of determining the weights and outputting the signal 625. At the beginning, using R ZZ The eigenvectors of the matrix to perform the initial replication. The port labeled "#1" is initialized with the eigenvector 631-1 corresponding to the largest singular value, the second port with the next eigenvector 631-2, and the Nth port at step 703 s Eigenvectors 631-N s To initialize the Nth s A port. These eigenvectors are guaranteed to be linearly independent and are generally the preferred values to begin operation. Alternatively, any substantially independent initial weight vector may be employed. For example, in an alternative embodiment, port # 1 is used with vector [100.. 0]' initialization, with port # 2 using the vector [0100.. 0]' initialize, and so on. After initialization in the individual ports, the method is carried out in this way for the single-user case of the flow chart of fig. 3 in the individual ports. That is, step 305 is a copy operation executed with an initial value. The resulting signal is processed in step 307Time offset correction (which includes sampling/interpolation if over-sampled initially) to generate substantially baud aligned samples is fed to a signal quality estimator which estimates the SINR substantially at the baud point in step 309. In step 311, it is determined for weight adaptationIt is either the CM method or the decision-directed method that is selected. If the SINR is below the predefined SINR threshold, then an optimization based on the partial characteristic recovery method (preferably the CM method) is performed in step 313 and the method returns to step 305 to start the next iteration for the weight vector most recently determined for this port. If the SINR is above the threshold, frequency offset correction is performed in step 315 and a single decision directed adaptation iteration is performed in step 317. It should be noted that if the time offset correction in step 307 is only approximate, a more accurate correction may be required for the decision-directed method, and such modifications will be appreciated by those skilled in the art. In the preferred embodiment, only one decision-directed iteration is performed. Alternatively, more than one decision-directed iteration is performed. When weights are determined, the weight vectors ultimately determined for each port are used for the replication operation and demodulation steps performed on the entire burst, since in the preferred embodiment only a subset of the samples in each burst is used. This replication operation in the preferred embodiment includes the determination and correction of time and frequency offsets and demodulation, preferably using the architecture shown in fig. 4. The result for each port is a demodulated signal 617.
It should be noted that one feature common to the present invention in both single-user and multi-user scenarios is the use of weights obtained with the current burst data to determine the signal of the current burst. When the subscriber unit is moving around and in other fading and low SINR environments, good results may not be obtained with the weight vectors obtained from the previous burst.
The final step is to sort the outputs to determine if any of the output ports has become locked by the interferer. A PHS burst (e.g., a traffic channel burst) includes fields for the payload, a Unique Word (UW) known to all subscriber units, and a Cyclic Redundancy Check (CRC) field for error detection. Other protocols include somewhat different fields that can be used to determine whether a particular message is from a particular subscriber unit or is intended for a particular base station. In the preferred embodiment, to determine interference lock, it is contemplated to distinguish a desired subscriber unit transmitting a waveform that is valid for the system from an interfering subscriber unit also transmitting a waveform that is valid for the system. That is, there are ways to define a "valid" subscriber unit waveform, e.g., having some desired data and modulation format, and scrambled with a particular key for that subscriber unit. Likewise, the interfering unit may include some means for defining its own validity, e.g., the waveform has some desired data and modulation format, and is scrambled with a particular key different from the subscriber unit. In the best PHS implementation, one method for detecting interference locking includes monitoring the Unique Word (UW) and CRC simultaneously. In particular, the data bits in each burst are scrambled in a bit pattern generated using the lower 9 bits of the cell station identification Code (CSID). These 9-bit words used to encrypt the burst payload and its associated CRC are referred to as scrambling keys. When designing a communication system, for example a cellular system, it may be advisable to ensure that adjacent communication stations (base stations) each have a different scrambling key. Note that in the PHS specification, a base station or a communication station is referred to as a cell station.
It is known that the basic interference lock detection method of the preferred embodiment comprises the following steps:
for a particular port, i.e. a particular subscriber unit
1. Demodulating the received signal with the reception right determined for the subscriber unit, descrambling the burst payload with a CSID based key for the subscriber unit;
2. comparing the received CRC with a CRC calculated from the demodulated descrambled bit sequence of the burst payload;
3. it is determined whether there is a significant difference between the two, indicating a transmission error or a secret key mismatch, while the UW does not show an error, and if the condition is met, a counter is triggered. In the case where the weight "track" is included, if the condition is not met, assuming that the communication is not interference locked, the weight (or subscriber unit spatial signature) received from the subscriber unit is saved ("tracked") as a "good" value for the subscriber unit; and
4. if a certain number of consecutive bursts satisfy the condition set forth in step 3, then the port is determined to be interference locked, as determined by the counter.
In the present PHS specification, when the co-channel users are all different spatial channels in the same conventional channel, using the unique word and CRC to determine interference locking will not work because the CSID is the same for all subscriber units of the same conventional channel. For the case where co-channel users are spatial channels of the same conventional channel, determining interference locking may be accomplished by maintaining the spatial signature history of these co-channel users.
Once the outputs are sorted, the result is a set of output signals from each port.
Device
Fig. 9 shows a block diagram of an apparatus implementing an aspect of the invention. Apparatus for determining a weight vector for receiving a particular signal transmitted by a particular subscriber unit, comprising: an initialization means 902 for initializing with a first initial vector value; first iteration means 905 for iteratively modifying the weight vectors according to a first iteration method minimizing a first cost function, which is an iteration weight determination method with good convergence properties, preferably a constant modulus method implemented as described above. The apparatus further comprises second iterative means 907 for iteratively modifying the weight vectors according to a second adaptive method, which minimizes a second cost function, which is a fast converging iterative weight determination method, preferably the decision-directed method described above. An initialization means 902, a first iteration means 905 and a second iteration means 907 under the control of the control means 911, said control means 911 being programmed to activate the first iteration means 905 starting from a first initial vector value provided by the initialization means 902 until a switching criterion is fulfilled, the final weight vector after the last iteration of said first adaptive method being the second vector value; and activating said second iteration means 907 starting from said second vector value to determine a weight vector 909. The weight vectors 909 are used by the spatial processor and demodulator 915 to generate demodulated signals at the direction of the controller 911, which spatial processor utilizes the signals received by the receiver 122 at the antenna array 103. Various iterative methods include determining a replica signal. The apparatus preferably includes a SINR estimator 913 for estimating a post-replica SINR in the first iteration means replica signal using the weight vector determined by the first iteration means 905, the switch criterion preferably being that the SINR estimate exceeds an SINR threshold.
The weight determining means preferably comprises at least one Digital Signal Processor (DSP) means at the base station, and each of the components 902, 905, 907, 909, 911, 913 and 915 is preferably implemented in the form of a program in one or more DSPs. It will be appreciated by those skilled in the art that many modifications may be made to the methods and apparatus described above without departing from the spirit and scope of the invention. For example, a communication station implementing the method may employ one of a plurality of protocols. In addition, a variety of architectures for these stations are possible. Many further variations are possible. The true spirit and scope of the present invention should be limited only by the terms set forth in the following claims.
Claims (33)
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/US2000/011563 WO2001081942A1 (en) | 1999-04-01 | 2000-04-26 | Adaptive smart antenna processing method and apparatus |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1454322A CN1454322A (en) | 2003-11-05 |
| CN100372404C true CN100372404C (en) | 2008-02-27 |
Family
ID=21741328
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB008196753A Expired - Fee Related CN100372404C (en) | 2000-04-26 | 2000-04-26 | adaptive smart antenna processing method and device |
Country Status (3)
| Country | Link |
|---|---|
| JP (1) | JP4674030B2 (en) |
| CN (1) | CN100372404C (en) |
| AU (2) | AU4679200A (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100965721B1 (en) * | 2003-06-30 | 2010-06-24 | 삼성전자주식회사 | Apparatus and method for receiving data in mobile communication system using adaptive antenna array method |
| US7260408B2 (en) * | 2004-02-20 | 2007-08-21 | Airespace, Inc. | Wireless node location mechanism using antenna pattern diversity to enhance accuracy of location estimates |
| WO2006092462A1 (en) * | 2005-03-01 | 2006-09-08 | Elektrobit System Test Oy | A method, device arrangement, transmitter unit and receiver unit for generating data characterising mimo environment |
| GB0505800D0 (en) * | 2005-03-22 | 2005-04-27 | Univ Sheffield | Control of processes |
| US7526048B2 (en) * | 2005-08-11 | 2009-04-28 | Mitsubishi Electric Research Laboratories, Inc. | Energy threshold selection for UWB TOA estimation |
| EP3864772A4 (en) * | 2018-10-09 | 2022-06-15 | Telefonaktiebolaget LM Ericsson (publ) | Method and apparatus for power control |
| CN114883803B (en) * | 2022-04-21 | 2025-01-03 | 中国空间技术研究院 | A microwave energy beam pointing control method for space solar power station |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5515378A (en) * | 1991-12-12 | 1996-05-07 | Arraycomm, Inc. | Spatial division multiple access wireless communication systems |
| CN1165458A (en) * | 1997-04-21 | 1997-11-19 | 北京信威通信技术有限公司 | Time division duplex synchronous code partition multi-address radio communication system |
| US5909470A (en) * | 1996-10-11 | 1999-06-01 | Arraycomm, Inc. | Method and apparatus for decision directed demodulation using antenna arrays and spatial processing |
| US5930243A (en) * | 1996-10-11 | 1999-07-27 | Arraycomm, Inc. | Method and apparatus for estimating parameters of a communication system using antenna arrays and spatial processing |
| US6018317A (en) * | 1995-06-02 | 2000-01-25 | Trw Inc. | Cochannel signal processing system |
| CN1286001A (en) * | 1997-12-10 | 2001-02-28 | 埃瑞康姆公司 | Radio transmission from communication station with antenna array to provide desirable radiation pattern |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3075983B2 (en) * | 1996-03-22 | 2000-08-14 | 株式会社ワイ・アール・ピー移動通信基盤技術研究所 | Receiver with adaptive antenna |
| US5887038A (en) * | 1997-04-22 | 1999-03-23 | Golden; Glenn David | Adaptive antenna array processing arrangement using a combined coherent and constant-modulus reference signal |
| JP3768350B2 (en) * | 1998-03-30 | 2006-04-19 | 松下電器産業株式会社 | Wireless receiving apparatus and method |
| JP2000106505A (en) * | 1998-09-28 | 2000-04-11 | Toshiba Corp | Adaptive array antenna |
-
2000
- 2000-04-26 JP JP2001578980A patent/JP4674030B2/en not_active Expired - Fee Related
- 2000-04-26 CN CNB008196753A patent/CN100372404C/en not_active Expired - Fee Related
- 2000-04-26 AU AU4679200A patent/AU4679200A/en active Pending
- 2000-04-26 AU AU2000246792A patent/AU2000246792B2/en not_active Ceased
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5515378A (en) * | 1991-12-12 | 1996-05-07 | Arraycomm, Inc. | Spatial division multiple access wireless communication systems |
| US6018317A (en) * | 1995-06-02 | 2000-01-25 | Trw Inc. | Cochannel signal processing system |
| US5909470A (en) * | 1996-10-11 | 1999-06-01 | Arraycomm, Inc. | Method and apparatus for decision directed demodulation using antenna arrays and spatial processing |
| US5930243A (en) * | 1996-10-11 | 1999-07-27 | Arraycomm, Inc. | Method and apparatus for estimating parameters of a communication system using antenna arrays and spatial processing |
| CN1165458A (en) * | 1997-04-21 | 1997-11-19 | 北京信威通信技术有限公司 | Time division duplex synchronous code partition multi-address radio communication system |
| CN1286001A (en) * | 1997-12-10 | 2001-02-28 | 埃瑞康姆公司 | Radio transmission from communication station with antenna array to provide desirable radiation pattern |
Non-Patent Citations (2)
| Title |
|---|
| CDMA系统中的自适应阵列天线技术. 张强.天津通信技术，第1期. 2000 * |
| CDMA系统中自适应阵列天线结合参数估计方法抗多址干扰技术的研究. 唐瑜，严梅，龚耀环，李乐民.通信学报，第20卷第7期. 1999 * |
Also Published As
| Publication number | Publication date |
|---|---|
| AU2000246792B2 (en) | 2005-06-09 |
| AU4679200A (en) | 2001-11-07 |
| JP2003532313A (en) | 2003-10-28 |
| CN1454322A (en) | 2003-11-05 |
| JP4674030B2 (en) | 2011-04-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US6177906B1 (en) | Multimode iterative adaptive smart antenna processing method and apparatus | |
| Winters | Signal acquisition and tracking with adaptive arrays in the digital mobile radio system IS-54 with flat fading | |
| CN100420166C (en) | Array receiver with subarray selection, method of use and receiver system incorporating same | |
| US7411547B2 (en) | Antenna array including virtual antenna elements | |
| US5875216A (en) | Weight generation in stationary interference and noise environments | |
| US6839573B1 (en) | Apparatus and method for beamforming in a changing-interference environment | |
| US5819168A (en) | Adaptive communication system and method using unequal weighting of interface and noise | |
| CA2248345C (en) | Adaptive array transmitter-receiver | |
| Cupo et al. | A four-element adaptive antenna array for IS-136 PCS base stations | |
| US20070127608A1 (en) | Blind interference mitigation in a digital receiver | |
| EP1667341A1 (en) | Method for calculating a weighting vector for an antenna array | |
| US6147985A (en) | Subspace method for adaptive array weight tracking | |
| US7369877B2 (en) | Multi-channel demodulation with blind digital beamforming | |
| WO2000079718A2 (en) | Null deepening for an adaptive antenna based communication station | |
| US7369631B1 (en) | Symbol synchronizer for software defined communications system signal combiner | |
| CN100372404C (en) | adaptive smart antenna processing method and device | |
| AU2000246792A1 (en) | Adaptive smart antenna processing method and apparatus | |
| KR100761420B1 (en) | Adaptive smart antenna processing method and apparatus | |
| US7047045B2 (en) | Symbol estimation-based decorrelator for estimating spatial signatures in a wireless communications system | |
| Maruta et al. | Multi modulus signal adaptation for semi-blind uplink interference suppression on multicell massive MIMO systems | |
| JP2004215171A (en) | Wireless communication device and method, and wireless communication system | |
| Maruta et al. | Enhanced semi-blind uplink interference suppression on multicell massive MIMO systems for multi modulus signals | |
| Takatori et al. | Smart antenna testbed for SDMA systems using STBC | |
| Mizusawa et al. | Spatial-Temporal Signal Processing for Broadband Wireless Systems | |
| Nunes | Mendes |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| ASS | Succession or assignment of patent right |
Owner name: A LEI YI COMMUNICATION CO.,LTD. Free format text: FORMER OWNER: ARRAYCOMM INC. Effective date: 20070413 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20070413 Address after: American California Applicant after: Arraycomm Inc. Address before: American California Applicant before: Arraycomm Inc. |
|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| ASS | Succession or assignment of patent right |
Owner name: INTEL CORP Free format text: FORMER OWNER: A LEI YI COMMUNICATION CO.,LTD. Effective date: 20090515 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20090515 Address after: American California Patentee after: Intel Corporation Address before: American California Patentee before: Arraycomm Inc. |
|
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20080227 Termination date: 20140426 |