MXPA00000489A - Determination of the length of a channel impulse response - Google Patents

Abstract

Techniques for adapting detection schemes used in receivers for receiving radio signals are described. The received signal is processed to determine, for example, an amount of time dispersion present in the radio channel. Based on this determination an appropriate detection scheme is selected for detecting the transmitted symbols. Various techniques for determining the dispersive or non-dispersive nature of the channel are described.

Description

DETERMINATION OF THE LENGTH OF A CHANNEL MPULSO RESPONSE BACKGROUND In recent years, wireless, digital, communications systems have been used to transport a variety of information across multiple locations. With digital communications, information is translated in digital or binary form, known as bits, for communications purposes. The transmitter maps this bitstream into a modulated symbol stream, which is detected at the digital receiver and mapped back into information bits. In wireless digital communications, the radio environment presents many difficulties that prevent successful communications, for example, those caused by multiple paths of the signal traversed by radio signals before reaching a receiver. A difficulty occurs when the multiple paths of the signal are of very different length. In this case, time dispersion occurs, in which multiple images of the signal arrive at the receiver's antenna at different times, giving rise to echoes of the signal. This causes intersymbol interference (ISI), where the echoes of a symbol interfere with the subsequent symbols.

The time dispersion can be mitigated using an equalizer. Common forms of equalization are provided by linear equalizers, decision feedback equalizers and sequence estimation equalizers, maximum likelihood (MLSE). A linear equalizer attempts to undo the effects of the channel by filtering the received signal. A decision feedback equalizer exploits the detection of previous symbols to cancel the intersymbol interference of the echoes of these previous symbols. Finally, an MLSE equalizer assumes different sequences of transmitted symbols and, with a dispersive channel model, determines the hypothesis that best fits the received data. These equalization techniques are well known to experts and can be found in normal textbooks such as J. G. Proa is, Digital Communications, 2nd ed. , New York: McGraw-Hill, 1989. In TDMA systems, such as D-AMPS and GSM, equalizers are commonly used. Of the three common equalization techniques, the MLSE equalization is preferable from an operational point of view. All sequences of possible transmitted symbols are considered in the MLSE equalizer. For each hypothetical sequence, samples of received signals are predicted using a multipath channel model. The difference between the samples of the received, predicted signals and the samples of the actual received signals, called the prediction error, provides an indication of how good a particular hypothesis is. The square magnitude of the prediction error is used as a metric to evaluate a particular hypothesis. This metric is accumulated for different hypotheses for use in determining the best hypotheses. This process is carried out efficiently using the Viterbi algorithm, which is a form of dynamic programming. However, under certain operating conditions, it is possible that signals arriving at a receiver do not create significant levels of intersymbol interference. When the ISI is negligible, or does not exist, the equalizer actually adds more noise to the detection statistics than it eliminates, particularly when the channel varies rapidly. Under these conditions, it would be desirable to turn off the equalizer in favor of another detection device, for example a differential detector, which may work better under dispersive conditions other than time. In addition, from the computational point of view, an equalizer is relatively complex compared to a differential detector. Thus, periodically turning off the equalizer in favor of a differential detector would save MIPS which, in turn, would reduce battery consumption. As another example, in direct sequence CDMA systems, RAKE receptors are commonly employed. However, if too many RAKE taps are used, operation is degraded. Accordingly, it would be desirable to offer a receiver in which a suitable detection technique can be identified and implemented dynamically, for example, a detector using a number of suitable channel branches.

COMPENDIUM According to the exemplary embodiments of the present invention, the characteristics of the radio channel are measured to determine a suitable detection strategy for the realization in a detector. For example, if the non-dispersive radio channel is determined, then it is possible to select a differential detector for operation as a symbol detector. Otherwise, if a time dispersive channel is detected, then an equalizer can be used to detect information symbols received in a receiver. In the same way, for: the CDMA, if the radio channel is non-dispersive, then a correlator detector can be selected. Otherwise, if a time dispersive channel is detected, then a RAKE receiver can be used. Different types of detector controllers may be established in accordance with the present invention to select a suitable detection scheme for a specific received signal. For example, a ratio of the received signal-to-noise parameters can be evaluated and compared with the threshold. Based on a result of the comparison, it is possible to implement an adequate detection scheme. For example, in a simple case, the comparison can indicate whether the channel is time-dispersive or non-time-dispersive. According to another of the exemplary embodiments, a specific number of channel branches that precisely model a particular radio channel can be identified and used to determine an appropriate detection scheme. According to other exemplary embodiments of the present invention, a ratio of an energy of a principal beam to the summed energies of the additional or secondary rays can be calculated to determine whether the channel is dispersive or non-dispersive. To avoid fluctuations due to fading, the energies can be weighted or smoothed before being compared to a threshold.

BRIEF DESCRIPTION OF THE DRAWINGS The features, objectives and advantages of the invention will be understood by reading the following detailed description together with the drawings, in which: FIGURE 1 is a block diagram illustrating ten cells in a mobile cellular radio system which invention is applied; FIGURE 2 is a general block diagram of a mobile station according to an aspect of the present invention; FIGURE 3 is a diagram illustrating a first exemplary embodiment of the detector controller illustrated in FIGURE 2; FIGURE 4 is a diagram illustrating a second exemplary embodiment of the detector controller of FIGURE 2; FIGURE 5 is a diagram illustrating a third exemplary embodiment of the detector controller of FIGURE 2; FIGURE 6 is a diagram illustrating a fourth exemplary embodiment of the detector controller of FIGURE 2; FIGURE 7 is a diagram illustrating "a fifth exemplary embodiment of the detector controller of FIGURE 2; and FIGURE 8 is a flow chart depicting an exemplary method for selecting an appropriate detection technique in accordance with an exemplary embodiment of the present invention.

DETAILED DESCRIPTION The different characteristics of the invention will now be described with respect to the figures, in which like parts are identified with the same reference characters. Although the following description is provided in the context of non-dispersed systems, those skilled in the art will appreciate that the present invention is equally applicable to dispersed systems (e.g., CDMA) as well. FIGURE 1 is a schematic diagram illustrating the relationship between 10 cells (C1-C10) in a common cellular telephone network 100 (hereinafter referred to as a "cellular network") such as D-AMPS. In general, a cellular network "would have much more than 10 cells, however, 10 is sufficient for illustrative purposes.In each Cl to CÍO cell there is a base station Bl to B10, although FIGURE 1 shows the base stations located towards the center of each cell, the base stations may be located anywhere in the cell.The base stations located toward the center usually employ omnidirectional antennas, while the base stations located toward a cell contour usually employ directional antennas.

The cellular network 100 shown in FIGURE 1 also has a mobile switching center (MSC). The MSC connects to each of the base stations by cable, radio links or both (not illustrated in FIGURE 1). The MSC is also connected to a fixed telephone switching unit (also not illustrated in FIGURE 1). The mobile M1-M10 represent mobile phone units. Of course, mobiles can move around a cell or can move from one cell to another cell. Usually, there are much more mobile than ten. Again, ten mobiles is sufficient for illustrative purposes. Each mobile station includes a receiver (also not illustrated in FIGURE 1) to receive signals transmitted over the air interface from a base station to which this mobile station is currently listening. The receiver processes the received information symbols, for example, using demodulation and detection techniques, to extract the information symbols included in the received signals. Usually, these receivers include a detection device, for example, an equalizer or a differential detector, used to identify the information symbols in the received signal flow. The selection of a particular detection device to include it in a receiver, -for example, an equalizer having some fixed, predetermined number of channel derivations, was usually made based on the radio environment in which it was proposed that the receiver to operate. The present invention, however, takes another approach. - ~ Referring to FIGURE 2, a general block diagram of a mobile station according to the present invention is illustrated. In this, in the antenna 20, a flow of received signals is received in a mobile station. This signal flow is then processed, eg, amplified, filtered and converted downward, in the radio receiver 22 according to the techniques known to produce a flow of complex baseband signal samples. The resulting flow is then fed to a detector controller 24 and a detector 26. The detector controller 24 processes the received signal flow, as will be described in more detail below, to determine an optimal technique for detecting the symbols of information in this flow. According to the results of this processing, the detector controller 24 will send an appropriate instruction to the detector 26, so that the detector 26 performs the selected detection technique. The output of the detector 26 is a stream of information symbols which is then processed downstream to send user information (eg voice or data) or to respond to the control information (eg, a location message). The manner in which it operates on detector controller 24 will now be described to select a particular detection technique. When a series of known synchronization symbols has been received, the receiver can then use the corresponding received data to form an output signal from the detector controller. For example, the synchronization symbols can be used to perform the estimation of the least squares channel or channel estimation using the correlations between the synchronization symbols and the received data. The channel estimation information can be used to model the radio channel as including J channel branches. For example, a Sest signal power can be estimated by adding the square magnitude of the channel branches, "A Sest (¿T) -? ¡Ci)! *, Where c (j) represents the estimates of the channel coefficient.At the same time, the estimates of the channel coefficient and known symbols can be used to form estimates of the data received, that is, rest (k) = c (0) s (k) + c (l) s (kl) + ... + c (Jl) s (k-J + l), where s (k) represents the known synchronization symbols These estimates of the data received can, in turn, be used to form iin estimated noise power, Nest (J), by averaging the square magnitude of r (k) -rest (k) over the synchronization data received. Thus, both Sest (J) and Nest (J) can be determined for different candidate values of J (for example J = l ... Jmax). Since the operation of the system is usually related to these quantities, it is possible to use a comparison device to determine how many channel derivations need to be modeled to provide a degree of system performance. Note that in the absence of known symbols ^ it is possible to use hypothetical symbols instead. Also, the number of derivations used may change over time (for example, within a TDMA time slot). Thus, the detector's controllers may vary the number of derivations used dynamically during reception. Having provided a conceptual overview of exemplary receptor structures in accordance with the present invention, various techniques and structures will now be described to determine a desired detection scheme. An exemplary embodiment of the detector controller 24 is shown in FIGURE 3, which is designed to determine whether or not it is dispersed (ie, if J is present). > 1 or J = 1, respectively). This information can then be used to select an appropriate detection mechanism. For example, for channels using DQPSK modulation, it is possible to select a differential detector when J = 1 and an equalizer can be selected when J > 1. In FIGURE 3, "the received data passes through the synchronization unit 30, which performs the synchronization for the purposes of demodulation assuming that the channel is non-dispersive, (ie, assuming that J = 1 as shows by the arrow that goes from the "l" to the block SYNC 30). The synchronized data is then used to determine an estimate of the single channel coefficient c (0) associated with a non-dispersive channel in the channel estimation unit 32, for example, using any of the aforementioned known techniques. The channel estimate and the synchronized data are used by the noise power estimator 34 to produce an estimate of the noise power over the synchronization field, called Nest (l). This can be done averaging I I r (k) -c (0) s (k) | I 2 on the synchronization field, since s (k), the synchronization symbols transmitted are known. The channel estimate also goes to the estimator of the power of the signal 36 that produces Sest (l), forming the square magnitude of the channel coefficient, c (0) Both Nest (l) and Sest (l) pass to a comparator 3Í which determines whether the signal-to-noise ratio exceeds a certain threshold T (whose threshold is determined by the minimum acceptable SNR for the proper functioning of the communications, which in turn can be determined by empirical evidence as will be appreciated by those skilled in the art) that is: Sest (1) / Nest (1) > T? This evaluation can be done in different ways to avoid division, such as comparing: Sest (l) > Nest (l) T? If the threshold is exceeded, then the detector collector 24 (Figure 2) sends a control signal to the controlled detector 26 indicating that it can use a non-dispersive signal detection form, eg, differential detection or coherent detection of the single derivation . Otherwise, the control signal indicates that a form of dispersive signal detection is necessary, for example, multi-derivative equalization. Thus, this exemplary embodiment of the detector controller 24 determines whether the channel is dispersive (J> 1) or not (J = 1). In FIGURE 4 a second exemplary embodiment of the detector controller 24 is illustrated. In this embodiment, the controller 24 determines how much dispersion is present. (that is, the value of J, the number of derivations of the channel). The synchronization 40 and estimation units of channel 42 both operate under the assumption that some maximum number of channel derivations are present.

(J ax), for example five derivations. Channel derivations and data are provided to a plurality of noise power estimators, one for each possible value of J. To simplify the drawings, only two of these estimators of noise power 44 and 46 are illustrated, is say, for derivations 1 and Jmax. For example, with each estimator assuming J derivations, the noise power is estimated using the difference between r (k) and c (0) s (k) + ... c (J-l) s (k-J + l). These estimates are compared with each other using the threshold T in the comparator 48. The value J is then determined so that: Nest (J + l) > T Nest (J) where T is a design parameter between 0 and 1 that can be determined empirically and usually is little less than 1, for example 0.9. This gives the number of channel derivations to be modeled in the detector 26. Note that this exemplary mode can be modified to further form Sest (J) (eg, as described with respect to FIGURE 3) for each value possible of J and use Nest (J) and Sest (J) in the comparator, that is: T Sest (J + l) / Nest (J + l) <; Sest (J) / Nest (J) which is equivalent to the comparison: Sest (J) Nest (J + l) > T Sest (J + l) Nest / J) A third exemplary mode of the detector controller is illustrated in FIGURE 5. Using the ellipse notation mark that is observed first in FIGURE 4 to define the additional branches that are not illustrated, to simplify the figure, only the blocks associated with J = 1 and J = Jmax are shown. Those skilled in the art will appreciate that similar branches would be provided for J = 2, 3, 4, ... etc. In this modality, different synchronization criteria are considered, each corresponding to different possible values for J, the number of the channel derivation coefficients. For example, the synchronization units 50 and 52 can be implemented to find the synchronization so that the energy in the first and the last (Jmax) of the channel coefficients, respectively, is maximized. Subsequent estimates of the channel in each branch (ie, performed in blocks 54 and 56) also assume J derivations, as do the noise power estimators, for example, blocks 58 and 60. Then, as in FIGURE 4, the estimates are compared in a similar way to that described in FIGURE 4. Again, it is possible to use the estimates of signal strength to improve performance in a manner similar to that described in FIGURE 4. 4. According to yet another exemplary embodiment of the invention, the energy ratio associated with the intersymbol interference (ISI) to the energy associated with the main beam can be used to estimate the amount. of dispersion by delay of a received signal. For example, assuming a channel model with derivation L: C (z) Co + Ciz + CL_? Z (L-l) then the delay can be estimated by evaluating the relationship: where Co is the coefficient of the channel associated with the first ray of the signal or the strongest and Ck is a matrix of the remaining channel coefficients. However, this relationship must be weighted or smoothed to take into account instantaneous variations of? associated with fading. This smoothing can be carried out using the information accumulated from the first estimates. For example, B? 0, * (a) »Y * 0, a (m-l) + a- ?? E0 m) f JB (a> «E, (« ~ 1. *. * < 1 ~ Y > £, { «), J2 (a) m) - * '* t jgn í-pt ) Ll where o (m) - = | CO £ Í?) | 2, E ^ m) =? JC 2 The value smoothed? (m), it can then be compared with a threshold to determine if the channel is dispersive or non-dispersive. As with the previous modalities, this information can then be used to select an appropriate detection technique. An exemplary embodiment is illustrated in FIGURE 6. In this, incoming complex samples are synchronized assuming a channel model with L-lead in block 64. The channel estimate, assuming L-leads, is performed by the unit estimate of the channel. channel 66 to determine the channel coefficients. The coefficient of the channel associated with the first ray or the strongest one passes to the function of the square of the quantity block 68. The remaining coefficients pass to other blocks of the square magnitude function (for example, blocks 70 and 72) whose results are they add up in the adder 74. The energies Co and ISI are smoothed, as already described by smoothing functions 76 and 78, respectively. the ratio of the smoothed energies is then compared to a threshold T in the comparator 80, whose result characterizes the channel as dispersive or non-dispersive. An alternative embodiment is illustrated in FIGURE 7, where like reference numbers are used to define similar devices. In this, the sum of the ISI energies is performed downstream of the smoothing functions 90 in the comparator 92, instead of immediately after the functions of the square quantity 88 as in FIGURE 6. This allows the determination of how much dispersion needs to be matched by forming? j for different values of J, such as: According to yet another exemplary embodiment, during the synchronization in block 64 the derivations L are not assumed. Instead, an iterative approach is taken in ~ where a channel derivation is assumed first and the process represented in FIGURE 6 is performed. or TIGURA 7. If the statistics of the resulting channel is acceptable based on the number of assumed channel derivations, then the process ends, otherwise, another iteration is performed assuming a different number of channel derivations. This technique is illustrated by means of the flow chart of FIGURE 8. In this, a variable of the derivation of the N channel is set to 1 for the first step in step 100. Then, in step 102, the ratio of the smoothed energy (ie, known as the "channel" statistics in FIGURE 8) based on the timing for the received signal assuming a channel derivation. If the channel statistic is greater than a threshold T, whose threshold is determined empirically to provide a suitable signal ratio to ISI, then N is an exact number of derivations to model this channel and the process is moved to step 106 where a signal is selected. suitable detection scheme for the detector 26. For example, if after the first iteration the channel statistics exceeds the threshold T, then the channel is non-dispersive and a differential detection scheme can be used. If, on the other hand, the channel statistics is less than the threshold T, then the flow moves to step 108. In this, the derivation variable of the N channel is incremented and the process is repeated assuming a model with a derivation of additional channel compared to the previous iteration for synchronization purposes. Nevertheless, note that the channel estimate is made based on the maximum number of derivations. The invention has been described with reference to a particular embodiment. However, it will be readily apparent to those skilled in the art that it is possible to incorporate the invention into specific forms in addition to the preferred embodiments described above. This can be done without departing from the spirit of invention. The preferred modalities are merely illustrative and should not be considered restrictive in any way. The scope of the invention is given by the attached clauses, instead of the preceding description, and all equivalent variations that fall within the range of the clauses are proposed to be contained therein.

Claims (17)

1. A receiver consisting of: processing circuits for receiving a radio signal and producing signal samples from them; a detector controller for evaluating the signal samples to produce a number of channel taps necessary for detection; and a detector, responsive to an output of the detector controller, for detecting the symbols using the resulting number of channel derivations.

2. The receiver of claim 1, wherein the detector uses a differential-detection scheme when the number resulting from the derivations of * channel is one and a multi-derivation equalization scheme when the resulting number of channel derivations is greater that one.

3. The receiver of claim 1, wherein the detector uses a coherent detection scheme for a single derivation when the resulting number of the channel derivations is one and a detection scheme of the RAKE receiver when the number resulting from the derivations of the channel is greater than one.

4. The receiver of claim 1 further comprising: estimating the power of the signal to estimate a signal strength associated with the received signal; and an estimator of the noise power to estimate a noise power associated with the received signal.

The receiver of claim 4, wherein the detector controller determines the number of derivations based on a comparison between the estimated signal power and the estimated noise power.

The receiver of claim 5, wherein the detector controller synchronizes for a received signal and estimates a channel associated with the received signal based on the assumption that the channel is non-dispersive.

The receiver of claim 1, wherein the detector controller determines the number of derivations based on a comparison of a plurality of noise power estimates from which a dispersion level is identified by time, in where each of the plurality of noise estimates is calculated assuming a different number of channel derivations.

The receiver of claim 7, wherein the detector controller synchronizes for the received signal and estimates a channel associated with the received signal assuming a maximum number of channel derivations, the channel estimates being used to calculate the estimates of the channel. noise power.

9. The receiver of claim 7 further comprises: an estimator of the signal strength associated with each of the different number of channel branches to estimate a signal power associated with the received signal, wherein the detector controller compares the ratios of the estimates of the signal power to the estimates of the noise power for each different number of channel branches with a threshold.

10. The receiver of claim 1, further comprising: a plurality of branches, each branch includes: a synchronization unit for synchronizing for the received signal assuming a predetermined number of branches of the channel; a channel estimation unit for estimating the channel coefficients associated with the received signal assuming the predetermined number of channel derivations; and a noise power estimation unit for estimating a noise power associated with the received signal assuming the predetermined number of channel shunts; wherein the predetermined number of branches of the channel differs for each of the plurality of branches; and a comparator to receive a result from each of the branches and compare the results to: determine the number of derivations.

11. The receiver of claim 10, wherein a result of a branch is compared to a threshold multiplied by a result of another branch to determine the number of derivations.

12. The receiver of claim 10 further comprises: a plurality of estimators of signal strength, each associated with the plurality of branches to estimate a signal power associated with the received signal, wherein the comparator compares a ratio of the estimate of the signal power to the estimate of the noise power for one of the plurality of signals. branches with a ratio of the estimate of the power of the signal to the estimate of the power of the noise for another of the plurality of branches multiplied by a threshold value.

The receiver of claim 1, wherein the detector controller produces the number of derivations based on a comparison of the energy associated with the intersymbol interference with the energy associated with a main beam of the received signal.

14. The receiver of claim 13, wherein the energies are smoothed before comparison.

15. A receiver consisting of: means for receiving a radio signal to produce a received signal; the means to process the received signal to produce complex samples; the means to process the complex samples to produce the estimates of the signal power and the estimates of the power of the noise; the means for comparing the estimates of the signal power and the estimates of the noise power to generate a control signal from the detector; and means for detecting digital symbols within the received signal using the complex samples and the control signal.

16. A method for selecting a number of channel derivations for a detection scheme in a receiver consists of the steps of: (a) initializing a variable number of N-channel derivations for one; (b) evaluate a channel statistic using the variable number of N channel derivations; (c) comparing the channel statistics with a threshold; (d) selectively select N as a number of channel derivations used in a detection scheme based on a result of the comparison step; and (e) otherwise, increase the variable number of derivations of the N channel and perform another iteration of the steps (b) - (e). The receiver of claim 1, wherein the detector controller produces the number of derivations based on a comparison of the energy associated with the intersymbol interference with the energy associated with the equalized rays of the received signal. _ _ __