DE102013112273A1 - Techniques for channel estimation - Google Patents

Abstract

A method of estimating multipath channel parameters comprises: receiving (701) a signal comprising reference signal transmissions over a multipath channel; Providing (702) the received signal in a first time domain representation; Detecting (703) a first set of signal peaks based on the first time domain representation of the received signal; Providing (704) a pool of multipath channel parameters based on the detected first set of signal peaks; Selecting (705) a particular set of multipath channel parameters from the pool of multipath channel parameters according to an optimality criterion.

Description

TECHNICAL AREA

The present disclosure described herein relates generally to channel estimation techniques. In particular, aspects of the present disclosure may relate to methods and apparatus for estimating multipath channel parameters, particularly methods and apparatus using a priori parametric channel estimation information, e.g. B. in OFDM systems (Orthogonal Frequency Division Multiplex).

STATE OF THE ART

Wireless communication networks may include multiple base stations and multiple mobile stations or user equipment (UE). Signals transmitted between components of a wireless communication network may be subject to multipath fading. Methods and devices used in wireless communication networks need to be constantly improved. In particular, it may be desirable to improve the channel estimation in wireless communication networks.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings are provided to provide a further understanding of aspects and are incorporated in and constitute a part of this specification. The drawings illustrate aspects and, together with the description, serve to explain principles of aspects. Other aspects and many of the intended advantages of aspects will be readily appreciated as they become better understood by reference to the following detailed description. Like reference numerals designate corresponding similar parts.

1 is a scheme of a wireless system 100 with a base station 110 and a mobile station 120 where the mobile station 120 Apply techniques for channel estimation.

2 is a scheme of a procedure 200 according to the disclosure, which uses an extended noise reduction algorithm (ENRA) as described below with respect to equation (3) for parametric channel estimation.

3 is a scheme of a procedure 300 according to the parametric channel estimation disclosure, the method including a PDP or power delay profile estimation part 320 and an iterative FBS and forward and backward selection estimating parts, respectively 340 includes.

4a FIG. 12 is a schematic of an exemplary performance delay profile 400a as it is after an accumulation step 322 of the PDP estimation part 320 of the procedure 300 as in 3 is provided mapped.

4b FIG. 12 is a schematic of an exemplary performance delay profile 400b as it did after an initial peak detection step 323 of the PDP estimation part 320 of the procedure 300 as in 3 is provided mapped.

4c FIG. 12 is a schematic of an exemplary performance delay profile 400c as it does after a resampling step 325 of the PDP estimation part 320 of the procedure 300 as in 3 is provided mapped.

5 is a scheme of an iterative FBD estimation part 340 a procedure 300 as in 3 shown for parametric channel estimation according to the disclosure.

6a is a diagram 600a an exemplary ETU performance profile 601 and an estimated performance profile 602 which are provided by using a parametric channel estimation method according to the disclosure.

6b is a performance diagram 600b a block error rate using a parametric channel estimation method according to the disclosure.

7 is a scheme of a procedure 700 for estimating multipath channel parameters according to the disclosure.

8th is a scheme of a procedure 800 for multipath channel parameter estimation according to the disclosure.

9 is a scheme of a channel estimator 900 according to the disclosure.

DESCRIPTION

In the following detailed description, reference is made to the accompanying drawings, which form a part hereof, and in which is shown by way of illustration specific aspects in which the disclosure may be practiced. It should be understood that other aspects may be utilized and structural or logical changes may be made without departing from the scope of the present disclosure. The following detailed description is therefore not to be considered in a limiting sense, and the scope of the present disclosure is defined by the appended claims.

OFDM:

Orthogonal-Frequenzmultiplex,

PCE:

parametrische Kanalschätzung,

PACE:

pilotunterstützte Kanalschätzung,

LTE:

Long Term Evolution,

CRS:

zellenspezifisches Referenzsignal,

ENRA:

erweiterter Rauschreduktionsalgorithmus,

ETU:

erweitertes typisches städtisches Kanalmodell

QAM:

Quadraturamplitudenmodulation,

TX:

Senden,

RX:

Empfangen,

IDFT:

inverse diskrete Fourier-Transformation,

MDL:

Minimalbeschreibungslänge,

ESPRIT:

Schätzung von Signalparametern durch Drehungsinvarianztechniken

SNR:

Rauschabstand,

PDP:

Leistungsverzögerungsprofil,

FBS:

Vorwärts-und-Rückwärts-Auswahl,

MSE:

mittlerer quadratischer Fehler,

MMSE:

minimaler mittlerer quadratischer Fehler,

HF:

Hochfrequenz,

UE:

Benutzergerät,

BLER:

Blockfehlerrate,

SINR:

Abstand von Signal zu Störungen und Rauschen

Tx_Div:

Sendediversität.

The following expressions, abbreviations and spellings are used:

OFDM:

Orthogonal frequency division multiplexing,

PCE:

parametric channel estimation,

PACE:

pilot-assisted channel estimation,

LTE:

Long Term Evolution,

CRS:

cell-specific reference signal,

ENRA:

extended noise reduction algorithm,

ETU:

extended typical urban canal model

QAM:

Quadrature amplitude modulation,

TX:

Send,

RX:

Receive,

IDFT:

inverse discrete Fourier transformation,

MDL:

Minimum Description Length,

ESPRIT:

Estimation of Signal Parameters by Rotation Invariance Techniques

SNR:

Noise ratio,

PDP:

Power delay profile

FBS:

Forward and reverse selection,

MSE:

mean square error,

MMSE:

minimal mean square error,

HF:

High frequency,

UE:

User device,

BLER:

Block error rate,

SINR:

Distance from signal to noise and noise

Tx_Div:

Transmit diversity.

The methods and devices described herein may be based on channel estimation, in particular, parametric channel estimation. It is understood that comments made in connection with a method described can also apply to a corresponding device which is designed to carry out the method, and vice versa. For example, when describing a specific method step, a corresponding device may comprise a unit for carrying out the method step described, even if such a unit is not expressly described or illustrated in the figures. It should also be understood that the features of the various exemplary aspects described herein may be combined with one another unless specifically stated otherwise.

The methods and devices described herein may be implemented in wireless communication networks, particularly 3G, 4G and CDMA based communication networks. The methods and devices described below may also be implemented in a base station (NodeB, eNodeB) or in a mobile device (or mobile station or user equipment (UE)). The devices described may include integrated circuits and / or passive elements and may be manufactured according to various technologies. For example, the circuits may be designed as logic integrated circuits, analog integrated circuits, mixed signal integrated circuits, optical circuits, memory circuits, and / or integrated passive elements.

The methods and devices described herein may be configured to transmit and / or receive radio signals. Radio signals may or may not include radio frequency signals radiated by a radio transmitter (or radio transmitter or sender) having a radio frequency in the range of about 3 Hz to about 300 GHz. The frequency range may correspond to frequencies of AC electrical signals used to produce and detect radio waves.

The methods and devices described herein may be used in radio receivers, e.g. Time domain receivers. A time domain receiver is a radio receiver designed to counteract the effects of multipath fading. This can be done by using multiple "sub-receivers" called taps, paths or fingers, that is, multiple correlators are each assigned to a different multipath component. Each tap or finger can independently decode a single multipath component. At a later stage, the contribution of all taps or fingers can be combined to best utilize the different transmission characteristics of each transmission path. This can result in higher signal-to-noise ratio (SNR) in a multipath environment. 1 shows a wireless system 100 with a base station 110 and a mobile station 120 where the mobile station 120 Channel estimation techniques as described below. The multipath channel through which a radio wave from a base station 110 to a mobile station 120 can propagate as the original (line of sight) wave pulse 101 by a number of multipath components 101 . 102 . 103 due to obstacles 112 . 113 be viewed sending. Multipath components are delayed copies of the original transmitted wave propagating through different echoes, each with a different size and arrival time in the receiver. Since each component contains the original information, if the size and time of arrival (phase) of each component in the receiver are calculated by a process called channel estimation, then all components can be coherently added to improve information reliability. In the 1 pictured three exemplary multipath components 101 . 102 and 103 can do different paths 201 . 202 . 203 correspond as in 4a . 4b and 4c is shown and described below.

The methods and devices described herein may be used for parametric channel estimation. Parametric channel estimation is an alternative approach to OFDM channel estimation on a pilot basis, such as Wiener filtering. PCE is a pilot-assisted channel estimation (PACE) method that can use the cell-specific reference signal (CRS) in a mobile network standard such as LTE. Parametric channel estimation can use the extended noise reduction (ENRA) algorithm, which is often superior to Wiener filtering, as long as it is given (almost) ideal channel propagation delay parameters, as in 6a and 6b shown below for a standard mobile network configuration such as LTE with ETU 10 MHz channel, 64-QAM, transmission mode 2 (TX diversity), normal cyclic prefix and 70 Hz Doppler spread. It is therefore desirable to estimate the multipath delay parameters with sufficient accuracy to take advantage of the extended noise reduction algorithm. The description of the ENRA begins with the channel transfer model: y _p = X _p T _p α + w _p (1)

Where X _{p is} the matrix of pilot symbols (pilot index p), y _{p is} the observed (received) pilot samples, w _{p is} the pilot noise samples, α is the vector of L complex-valued channel tap nesting factors, and T _{p is} the delay transformation matrix representing the set of propagation delays for the L channel taps :

Here, fp (s) the frequencies of the pilot sub n, T _s is the sampling period, τ _l is the delay of the l th channel tap, that is, the delays of the power delay profile, N _p is the number of the pilot carriers, and N the DFT length.

Based on the knowledge of T _p , the auto covariance matrix P = E {αα ^H } of channel tap gain factors α, ie, the power delay profile and noise power gain powers σ 2 / w For example, the ENRA algorithm may generate the MMSE estimate of path gain factors as follows: α ^ = PT H / p (T _p PT H / p + σ 2 / wI _M ) ^-1 y _p (3)

The ENRA estimate (3) may thus require the estimation of three parameters, namely, the covariance matrix P, the number of channel taps L, and the amount of time delays {τ} of each channel tap. After these parameters have been estimated, time / frequency interpolation can be performed to estimate the channel frequency response at each frequency. Frequency interpolation can be obtained by applying a delay transformation matrix based on the estimated delays to α ^. Time interpolation can be performed directly on α ^ using techniques such as linear interpolation.

2 is a scheme of a procedure 200 according to the parametric channel estimation disclosure using the parametric ENRA channel estimation as described by equation (3). A channel parameter estimation block 201 may receive a signal comprising demodulated CRS pilot signals, ie received pilot signals after demodulation in the receiver. The channel parameter estimation block 201 For example, a number of channel taps L and a set of time delays {τ} of each channel tap, e.g. B. calculated according to a method 300 as below with reference to 3 described, calculate. A covariance matrix generation block 202 may calculate a covariance matrix P based on the number of channel taps L. A noise power estimation block 203 can the noise power σ 2 / w to calculate. A delay transformation matrix calculation block 204 may calculate the delay transformation matrices {T} on the basis of the number of channel taps L and the amount of time delays {τ}, e.g. B. according to equation (2). An MMSE calculation block 205 may determine the minimum mean square error of the path gain factors, ie the channel tap gain factors α, based on the noise power σ 2 / w, the delay transformation matrices {T} and the covariance matrix P z. According to equation (3), and can output the path gain factors α ^ with the minimum mean squared error. A time / frequency interpolation block 206 may interpolate the obtained MMSE path gain factors α ^ with respect to time and / or frequency to obtain the estimated channel parameters and thus the channel estimate.

In one example, the computation of the covariance matrix P may be simplified by assuming a uniformly distributed power delay profile such that P = E {αα ^H } = siag (σ 2/1, ..., σ 2 / L) = diag (1 / L, ..., 1 / L).

In the channel estimator can be an estimate σ ^ 2 / w the noise power available. Then the channel parameter estimation block 201 can be simplified, since only the multipath delay parameters (L, τ) may need to be calculated.

In one example, the channel parameter estimation block 201 the two blocks of the PDP estimate 320 and iterative FBS 340 as below with reference to 3 described. The PDP estimation block 320 may perform an Inverse Discrete Fourier Transform (IDFT) to obtain an initial indication of the channel parameters (L, τ). The iterative FBS block 340 can use a forward and backward selection algorithm to refine the results of the IDFT and select the channel taps giving near optimal performance.

3 is a scheme of a procedure 300 according to the parametric channel estimation disclosure, the method including a PDP or power delay profile estimation part 320 and an iterative FBS and forward and backward selection estimating parts, respectively 340 includes.

In the PDP or power delay profile estimation part 320 For example, an estimate of the power delay profile (PDP) may be generated by transforming sets of the received CRS pilot signals, ie, the demodulated CRS pilot signals in the time domain, e.g. By applying an inverse discrete Fourier transform over an IDFT block 321 , The absolute square values of the IDFT output in each subframe according to a mobile network standard such as LTE may be e.g. Calculated by non-coherent estimation and then in an accumulation block 322 be accumulated or averaged over a number M of subframes. By means of this averaging, the noise level can be suppressed and the results are more reliable. In a beginning peak detection block 323 then a threshold can be applied, for. By applying truncation to separate the "real" peaks of noise. The peaks detected by the clipping applied by the preamble providing block 324 may provide an initial indication of the number of channel taps L ₀ and their delays {τ} ₀ . The clipping can be applied adaptively: starting with zero, the threshold can be implemented by a small step, after each increment the number of detected peaks be counted. This incrementing can be continued until there is a sudden decrease in the detected number of peaks, thus indicating the transition from the noise level region to the signal region.

The over the IDFT block 321 generated PDP estimate can only have a rough resolution of delays. To approach the true tap delays more closely, a refinement block may be provided 325 perform a refinement in various ways for each of the parameters L, and τ: The number of channel taps L can be expanded by adding more possible settings to the IDFT results, e.g. For example, there are four additional taps that can give a pool of channel tap numbers according to L = [L _IDFT , L _IDFT + 1, ..., L _IDFT + 4]. The channel path delays τ may be sampled at a particular granularity to encompass or closely approach the ideal values. Sampling can be performed on the time lags of the detected peaks (via IDFT) with a small step size, e.g. 10 nanoseconds. This type of sampling can be limited to a subset of delays, e.g. For the peaks detected in the first 1000 nanoseconds, where most channel taps are expected, thus reducing the delay of the path delay selection algorithm as described below.

The sampled values for the number of channel taps L and the channel path delays τ as provided by the initial pool of channel taps and the channel path delays 341 allow the iterative FBS (forward and backward selection) estimation block 342 which may execute a selection algorithm, such as those referred to below with reference to FIG 5 described "forward and backward selection" (FBS) algorithm to find the channel parameters that give near optimal performance.

The "forward and backward selection" or FBS algorithm, as determined by the FBS estimation block 342 after each recently added path delay can check for and modify the least significant previously selected path delays to approach optimal performance. The optimization or the optimality criterion may be, for example, the minimum mean square error (MSE). The FBS selection algorithm may be referred to as a bottom-up approach because it can begin with the assumption of only one path and then increment the number of paths step by step until near-optimal performance is achieved. If the channel is sparse, ie, has a small number of significant paths, the bottom-up approach may require fewer iterations than the opposite top-down approach, which may take many channel paths and may continue to discard them a certain optimization or optimality criterion is met. In the special case of an AWGN channel, the bottom-up approach can converge very quickly, and this can result in a much shorter runtime. The resulting estimates may be through a pool of end channel estimates and channel path delays 343 for further processing, eg. By the ENRA block 360 providing an ENRA algorithm. Another example of the implementation of the ENRA algorithm can be found in " Baoguo Yang; Letaief, KB; Cheng, RS; Zhigang Cao: Channel estimation for OFDM transmission in multi-path fading channels based on parametric channel modeling, Communications, IEEE Transactions on, Volume 49, No. 3, pp. 467-479, March 2001 ".

4a FIG. 12 is a schematic of an exemplary performance delay profile 400a as it is after an accumulation step 322 of the PDP estimation part 320 of the procedure 300 as in 3 is provided mapped. The diagram shows the resulting power delay profile 400a as the results of the IDFT and averaging over 100 subframes for an ETU-70Hz-64QAM channel model defined by LTE standardization.

4b FIG. 12 is a schematic of an exemplary performance delay profile 400b as it did after an initial peak detection step 323 of the PDP estimation part 320 of the procedure 300 as in 3 is provided mapped. The diagram shows the initial peak detection for the power delay profile 421 that the PDP profile 400a as in 4a shown corresponds. The peak detection block may be the channel tips 422 separate from noise. As 4b shows, the results after detection of the peaks can be eight channel taps 422 with delays τ = [0.0130, 0.0280, 0.0540, 0.0720, 0.1650, 0.2150, 0.2300, 0.5060] × 10 ^-5 (s). These channel taps or peaks can be used as an initial indication of the propagation channel delay parameters.

4c FIG. 12 is a schematic of an exemplary performance delay profile 400c which after a refinement step, e.g. A resampling step 325 of the PDP estimation part 320 of the procedure 300 as in 3 shown, is provided. The diagram shows the scan with a granularity of 10 Nanoseconds in a 1000 nanosecond window for the channel tips 431 that the in 4b represented channel tips 422 can correspond. Starting from these channel tips 422 For example, the number of channel taps may be extended by four more values (ie, L = [8, 9, 10, 11, 12]), and the path delays detected in a 1000 nanosecond window may be as in the diagram 400c be scanned shown. The resulting pool 432 , which contains the number of channel taps L and path delays τ, can then be forwarded to the "FBS forward and backward" algorithm, e.g. As by the FBS block 342 of the iterative FBS part 340 implemented as in the above 3 shown to estimate the number of channel taps and path delays. How out 4c In this example, the algorithm results in 10 channel taps with delays near the true path delays. The performance of each tap is not considered in the performance profile figures because the diagonal auto covariance matrix P is replaced by the uniformly distributed assumption of the diagonal matrix with the entries 1 / L.

5 is a scheme of an iterative FBD estimation part 340 a procedure 300 as in 3 as shown in the disclosure for parametric channel estimation. The iterative FBS estimation part 340 can by a forward-and-backward selection or FBS algorithm 500 as described below. The exemplary implementation of the forward-and-backward selection is in FIG 5 displayed.

The FBS algorithm 500 can begin iterating through the input pools of L, and τ. It can start with the first channel tap count L in the pool, as by block 501 and then iterate in the pool of path delays τ, as by block 504 and the blocks 507 and 508 mapped to select the path delay for each channel tap.

It can be a stop criterion 502 . 506 be applied to limit the number of iterations and also to solve the problem of MSE accuracy. During iterations looking for the best combination of path delays, the MSE can be calculated each time per iteration, as by block 505 displayed. The process can be stopped when the MSE starts to be stable or increases.

Thereafter, the selection algorithm may go to the next assumption 501 to skip channel tap counts. The FBS selection algorithm may interrupt 503 when a specified MSE accuracy is achieved, as by block 502 shown.

With regard to acquisition or block-wise tracking, the selection algorithm may be executed once every M subframe after the IDFT results may have been accumulated over M subframes to suppress the noise and to decrease the runtime of the search algorithm. The thus found channel parameters can be used for ENRA estimation z. In accordance with equation (3) during the next M subframes while a new round of PDP estimation is being performed, i. H. IDFT and accumulation during this new M subframe.

6a is a diagram of an example ETU power profile 601 and an estimated performance profile 602 which are provided by using a parametric channel estimation method according to the disclosure. The performance checks are configured with the parameters as shown in Table 1. bandwidth 10 MHz modulation 64-QAM code rate 3.4 antenna configuration 2 TX, 2 RX channel model EPA, ETU transfer mode 2 (Tx_Div) Doppler frequency 5 Hz SINR range [-10: 5: 20] Pool of channel tap delays calculated by IDFT Table 1: Configuration parameters of the performance measurement

As can be seen from the figure, the number of channel taps and path delays 602 the ideal profile 601 Near.

6b FIG. 10 is a performance diagram of a block error rate using a parametric channel estimation method in accordance with the disclosure. FIG. The "PCE-appreciated" 614 The disclosed algorithm is obviously (robust) Wiener filtering 612 superior and parametric channel estimation (PCE) 613 with ideal knowledge of channel parameters similar. The perfect channel estimation curve 611 can be used as a reference curve.

7 is a scheme of a procedure 700 for estimating multipath channel parameters according to the disclosure. The procedure 700 may comprise receiving a signal comprising reference signal transmissions over a multipath channel according to a first block 701 , The procedure 700 may provide the received signal in a first time domain representation according to a second block 702 include. The procedure 700 may comprise detecting a first set of signal peaks based on the first time domain representation of the received signal according to a third block 703 include. The procedure 700 may provide a pool of multipath channel parameters based on the detected first set of signal peaks according to a fourth block 704 include. The procedure 700 may include selecting a particular set of multipath channel parameters from the pool of multipath channel parameters according to an optimality criterion according to a fifth block 705 include.

The procedure 700 can use channel parameter estimation according to the in 2 pictured block 201 to implement. The received signal may correspond to the signal comprising the demodulated CRS pilot signals, as in FIG 2 and 3 displayed. Providing the first time-domain signal representation 702 can the blocks 321 and 322 of the PDP estimation part 320 as in 3 shown correspond. Detecting the first set of signal peaks 703 can the blocks 323 and 324 of the PDP estimation part 320 as in 3 shown correspond. Providing the pool of channel parameters 704 can the block 325 of the PDP estimation part 320 as in 3 shown correspond. Selecting the specific set of channel parameters 705 can be the iterative FBS part 340 as in 3 shown correspond.

In an example of the method 700 For example, providing the received signal in the first time domain representation may comprise applying an inverse discrete Fourier transform to the received signal. In an example of the method 700 For example, the reference signal transmissions may comprise orthogonal frequency division multiplexed transmissions of a cell specific reference signal, e.g. According to a mobile network standard such as LTE. In an example of the method 700 For example, detecting the first set of signal peaks may be based on an adaptive threshold crossing of an average power spectrum of the first time domain representation of the received signal. In an example of the method 700 For example, providing the pool of multipath channel parameters may include providing the received signal in a second time domain representation. A granularity of the received signal in the second time domain representation in a range of the signal peaks of the first set of signal peaks may be higher than a granularity of the received signal in the first time domain representation. In an example of the method 700 For example, providing the pool of multipath channel parameters may comprise detecting a second set of signal spikes based on a threshold crossing of an average power spectrum of the second time domain representation of the received signal. In an example of the method 700 For example, providing the pool of multipath channel parameters may comprise providing the pool of multipath channel parameters based on the detected second set of signal peaks. In an example of the method 700 For example, the pool of multipath channel parameters may include channel taps and respective delay values associated with the channel taps. In an example of the method 700 For example, selecting the particular set of multipath channel parameters may include starting from an initial number of channel taps and increasing the initial number in an iterative manner until a stop criteria is met. In an example of the method 700 For example, selecting the particular set of multipath channel parameters may include selectively adding and removing channel taps of the pool of multipath channel parameters to the determined set of multipath channel parameters for each iteration. In an example of the method 700 For example, the optimality criterion may be based on an extended noise reduction algorithm, e.g. As described above with reference to Eqs. (1) to (3).

8th is a scheme of a procedure 800 for multipath channel parameter estimation according to the disclosure. The procedure 800 receiving a signal comprising at least one demodulated cell-specific reference signal according to the first block 801 include. The procedure 800 the providing of a first and a second averaged power spectrum may comprise a time domain transformation of the received signal, wherein a granularity of the second averaged power spectrum in a range of signal peaks of the first averaged power spectrum is higher than a granularity of the first averaged power spectrum, according to a second block 802 , The procedure 800 For example, providing a pool of multipath channel parameters comprising channel taps and associated delay values may be based on peaks of the second average power spectrum according to a third block 803 include. The procedure 800 may determine a set of multipath channel parameters from the pool of multipath channel parameters based on an optimality criterion according to a fourth block 804 include.

In an example of the method 800 can the procedure 800 comprising determining candidate sets of multipath channel parameters based on forward and backward selection by the pool of multipath channel parameters. In an example of the method 800 can the procedure 800 performing the forward and backward selection from an initial number of channel taps in an iterative manner. In an example of the method 800 can the procedure 800 selecting combinations of channel taps from the pool of multipath channel parameters according to a current iteration. In an example of the method 800 can the procedure 800 determining mean squared errors between the at least one demodulated cell specific reference signal and path gain estimates of the respective candidate sets of multipath channel parameters. In an example of the method 800 can the procedure 800 performing the forward and backward selection by the pool of multipath channel parameters such that the mean square errors decrease. In an example of the method 800 can the procedure 800 interrupting forward and backward selection include when a decrease in mean square errors exceeds a threshold. In an example of the method 800 can the procedure 800 determining the path gain estimates by using parametric channel estimation based on a respective candidate set of multipath channel parameters. In an example of the method 800 can the procedure 800 determining the path gain estimates by using the following parameters: a delay transformation matrix representing propagation delays for the channel taps of a respective candidate set of multipath channel parameters; an autocovariance matrix of channel tap gain factors of the respective candidate set of multipath channel parameters; and a noise power of the at least one demodulated cell-specific reference signal.

9 is a scheme of a channel estimator 900 according to the disclosure. The channel estimator 900 can be a first unit 901 for receiving a signal 902 comprising at least one demodulated cell-specific reference signal. The channel estimator 900 can be a second unit 903 to provide a pool of multipath channel parameters 906 based on the received signal 904 include. The channel estimator 900 can be a third unit 905 for determining a set of multipath channel parameters 908 from the pool of multipath channel parameters 906 based on an optimality criterion.

The channel estimator 900 can be used to implement the above with reference to 7 described method 700 be used. The channel estimator 900 can also be used to implement the above with reference to 8th described method 800 be used. The channel estimator 900 can be used to implement the above with reference to 2 described method 200 be used. The channel estimator 900 can be used to implement the above with reference to 3 described method 300 be used.

In one example, the third unit 905 for determining a set of multipath channel parameters 908 may be used based on a mean squared error between the at least one demodulated cell specific reference signal and a result of one of a candidate set of multipath channel parameters from the pool of multipath channel parameters 906 applied extended noise reduction algorithm is calculated.

The present disclosure presents a reliable algorithm for estimating the multipath propagation delay parameters necessary for parametric channel estimation. The channel estimation techniques presented here may perform PDP estimation via IDFT to give an initial indication of the number of channel taps and their delays, thereby simplifying the task of the subsequent selection algorithm. The forward-and-backward FBS algorithm presented here gives reliable estimation performance for different channel types. The channel estimation techniques presented here do not depend on eigenvalues and therefore show reduced ones computational complexity. Performance is stable and reliable for a wide range of channel profiles and SNR values.

The present disclosure presents a robust and accurate technique for estimating channel parameters required as input by ENRA channel estimation. The channel estimation techniques presented here have low complexity because no eigenvalue decomposition or spin invariance techniques are required. The channel estimation techniques presented here are not sensitive to signal to noise ratio (SNR) variations; even at low SNR, the paths can be detected with smaller powers overwhelmed by noise. the performance is robust, which makes realistic applications possible. The channel estimation techniques presented here are of very low computational complexity and achieve stable performance even with fluctuating SNR. The number of channel taps L and the path delay τ for each tap are robustly and accurately estimated insensitive to errors in channel parameters, even errors in the channel path delays τ.

EXAMPLES

The following examples relate to further embodiments. Example 1 is a method of estimating multipath channel parameters, the method comprising: receiving a signal comprising reference signal transmissions over a multipath channel; Providing the received signal in a first time domain representation; Detecting a first set of signal peaks based on the first time domain representation of the received signal; Providing a pool of multipath channel parameters based on the detected first set of signal peaks; Selecting a particular set of multipath channel parameters from the pool of multipath channel parameters according to an optimality criterion.

In Example 2, the subject matter of Example 1 may optionally include providing the received signal in the first time domain representation comprising applying an inverse discrete Fourier transform to the received signal.

In Example 3, the subject matter of any one of Examples 1-2 may optionally include the reference signal transmissions comprising orthogonal frequency division multiplex encoded transmissions of a cell specific reference signal.

In Example 4, the subject matter of any one of Examples 1-3 may optionally include detecting the first set of signal peaks based on adaptive threshold crossing of an average power spectrum of the first time domain representation of the received signal.

In Example 5, the subject matter of any one of Examples 1-4 may optionally include providing the pool of multipath channel parameters comprising: providing the received signal in a second time domain representation, wherein a granularity of the received signal in the second time domain representation is in a range of the signal peaks first set of signal peaks is higher than a granularity of the received signal in the first time domain representation; Detecting a second set of signal peaks based on threshold crossing of an average power spectrum of the second time domain representation of the received signal; and providing the pool of multipath channel parameters based on the detected second set of signal peaks.

In Example 6, the subject matter of any of Examples 1-5 may optionally include the pool of multipath channel parameters comprising channel taps and respective delay values associated with the channel taps.

In Example 7, the subject matter of Example 6 may optionally include selecting the particular set of multipath channel parameters: starting from an initial number of channel taps and increasing the initial number in an iterative manner until a stop criteria is met.

Optionally, in Example 8, the subject matter of any one of Examples 6-7 may include selecting the particular set of multipath channel parameters: selectively adding and removing channel taps of the pool of multipath channel parameters to the determined set of multipath channel parameters for each iteration.

In Example 9, the subject matter of any of Examples 1-8 may optionally include the optimizer criterion being based on an enhanced noise reduction algorithm.

Example 10 is a method for multipath channel parameter estimation, the method comprising: receiving a signal comprising at least one demodulated cell-specific reference signal; Providing a first and a second averaged power spectrum of a time domain transform of the received signal, wherein a granularity of the second averaged power spectrum in a range of signal peaks of the first averaged power spectrum is higher than a granularity of the first averaged power spectrum; Providing a pool of multipath channel parameters comprising channel taps and associated delay values based on peaks of the second averaged power spectrum; and determining an amount of multipath channel parameters from the pool of multipath channel parameters based on an optimality criterion.

In Example 11, the subject matter of Example 10 may optionally include determining candidate sets of multipath channel parameters based on forward and backward selection by the pool of multipath channel parameters.

In Example 12, the subject matter of Example 11 may optionally include iteratively performing forward and backward selection from an initial number of channel taps.

In Example 13, the subject matter of Example 12 may optionally include selecting combinations of channel taps from the pool of multipath channel parameters according to a current iteration.

In Example 14, the subject matter of any of Examples 1-13 may optionally include determining mean squared errors between the at least one demodulated cell specific reference signal and path gain estimates of respective candidate sets of multipath channel parameters.

In Example 15, the subject matter of Example 14 may optionally include performing forward and backward selection by the pool of multipath channel parameters such that the mean squared errors decrease.

In Example 16, the subject matter of any of Examples 14-15 may optionally include interrupting the forward and reverse selections when a decrease in mean square errors exceeds a threshold.

In Example 17, the subject matter of any of Examples 14-16 may optionally include determining the path gain estimates by using parametric channel estimation based on a respective candidate set of multipath channel parameters.

In Example 18, the subject matter of any of Examples 14-17 may optionally include determining the path gain estimates by using the following parameters: a delay transformation matrix representing propagation delays for the channel taps of a respective candidate set of multipath channel parameters; an autocovariance matrix of channel tap gain factors of the respective candidate set of multipath channel parameters; and a noise power of the at least one demodulated cell-specific reference signal.

Example 19 is a channel estimator, comprising: a first unit configured to receive a signal including at least one demodulated cell-specific reference signal; a second unit configured to provide a pool of multipath channel parameters based on the received signal; and a third unit configured to determine a set of multipath channel parameters from the pool of multipath channel parameters based on an optimality criterion.

In Example 20, the subject matter of Example 19 may optionally include the third unit configured to determine the set of multipath channel parameters based on a mean squared error between the at least one demodulated cell specific reference signal and a result of a candidate set is calculated from multipath channel parameters from the pool of multipath channel parameters, applied extended noise reduction algorithm.

Example 21 is a receiver comprising: means for receiving a signal comprising at least one demodulated cell-specific reference signal; Means for providing a pool of multipath channel parameters based on the received signal; and means for determining a set of multipath channel parameters from the pool of multipath channel parameters based on an optimality criterion.

In Example 22, the subject matter of Example 21 may optionally include the optimizer criterion being based on an ENRA algorithm.

In Example 23, the subject matter of Example 21 or 22 may optionally include the optimality criterion being based on a minimum mean squared error.

In Example 24, the subject matter of Example 23 may optionally include computing the mean square error between a received reference signal, in particular a received CRS pilot signal, and results of an ENRA algorithm.

In Example 25, the subject matter of any of Examples 21-24 may optionally include the optimality criterion being based on a forward and backward selection.

In Example 26, the subject matter of any of Examples 21-25 may optionally include the optimality criterion being based on a forward selection.

Example 27 is a transmission system comprising a receiver according to any of Examples 21-26; and a transmitter adapted to transmit a radio signal comprising a cell-specific reference signal over a plurality of antenna ports.

In Example 28, the subject matter of Example 27 may optionally include a plurality of transmit antennas coupled to the antenna ports of the transmitter; and a plurality of receiving antennas coupled to the antenna ports of the receiver.

Example 29 is a base station comprising: means for receiving a signal comprising at least one demodulated reference signal; Means for providing a pool of multipath channel parameters based on the received signal; and means for determining a set of multipath channel parameters from the pool of multipath channel parameters based on an optimality criterion.

In Example 30, the subject matter of Example 29 may optionally include that the optimality criterion is based on an ENRA algorithm.

Example 31 is a mobile station comprising: means for receiving a signal comprising at least one demodulated cell-specific reference signal; Means for providing a pool of multipath channel parameters based on the received signal; and means for determining a set of multipath channel parameters from the pool of multipath channel parameters based on an optimality criterion.

In Example 32, the subject matter of Example 31 may optionally include the optimizer criterion being based on an ENRA algorithm.

In example 33, the subject matter of example 31 or 32 may optionally include the means for providing a pool of multipath channel parameters comprising a unit configured to process a frequency-in-time transform, in particular an inverse discrete Fourier transform, in particular an inverse fast Fourier Transformation on the received signal.

In Example 34, the subject matter of any of Examples 31-33 may optionally include encoding the cell specific reference signal according to an orthogonal frequency division multiplexing technique.

Example 35 is a method for multipath channel parameter estimation, the method comprising: receiving a signal comprising at least one demodulated cell-specific reference signal; Providing a first averaged power spectrum of a time domain transform of the received signal; Providing a pool of multipath channel parameters comprising channel taps and associated delay values based on signal peaks of the first averaged power spectrum; and Determining a set of multipath channel parameters from the pool of multipath channel parameters based on an optimality criterion.

In Example 36, the subject matter of Example 35 may optionally include determining candidate sets of multipath channel parameters based on forward and backward selection by the pool of multipath channel parameters.

In example 37, the subject matter of example 36 may optionally include iteratively performing the up-and-down selection of an initial number of channel taps.

In Example 38, the subject matter of any of Examples 35-37 may optionally include providing the signal peaks of the first averaged power spectrum based on a first adaptive threshold.

In example 39, the subject matter of example 38 may optionally include adjusting a step size of the first adaptive threshold such that the step size is increased when noisy power peaks are detected and decreased when stable power peaks are detected.

In Example 40, the subject matter of Example 39 may optionally include resampling the received signal using a higher sampling rate to provide a second power spectrum.

In example 41, the subject matter of example 40 may optionally include providing signal peaks of the second power spectrum based on a second adaptive threshold.

In Example 42, the subject matter of Example 41 may optionally include adjusting a step size of the second adaptive threshold based on the first adaptive threshold.

In Example 43, the subject matter of Example 42 may optionally be adapted to adjust a second adaptive threshold step size starting from a first adaptive threshold, increasing the step size when noisy second power spectrum power peaks, and decreasing the step size when using second power spectrum stable power peaks be detected include.

In Example 44, the subject matter of Example 43 may optionally include determining the amount of multipath channel parameters based on the detected peak power of the first power spectrum or the detected peak power of the second power spectrum.

In Example 45, the subject matter of Example 44 may optionally include determining the amount of multipath channel parameters using an ENRA algorithm.

In Example 46, the subject matter of any of Examples 35-45 may optionally include providing the pool of multipath channel parameters based on determining path gain estimates by using the following parameters: a delay transform matrix of the delay values of the channel taps; an autocovariance matrix of the channel taps; and a noise power of the at least one demodulated cell-specific reference signal.

In Example 47, the subject matter of Example 46 may optionally include the autocovariance matrix comprising a diagonal matrix, in particular a unitary matrix of uniformly distributed power delay profile.

Example 48 is a method for determining a set of multipath channel parameters from a pool of multipath channel parameters, the method comprising: determining candidate sets of multipath channel parameters based on forward and backward selection by the pool of multipath channel parameters; Determining mean squared errors of the candidate sets in an iterative manner; and discarding candidate sets for which the associated mean square error increases.

In Example 49, the subject matter of Example 48 may optionally include determining the candidate sets of multipath channel parameters based on an ENRA algorithm.

Example 50 is a computer-readable medium having stored thereon computer instructions that, when executed by a computer, cause the computer to perform the method of any one of Examples 1-18 and 35-49.

In addition, while a particular feature or aspect of the disclosure may have been disclosed with reference to only one of several implementations, such feature or aspect may be combined with one or more other features or aspects of the other implementations as may be substituted for one any given or specific application is desirable and advantageous. Further, where the terms "including," "comprising," "having," or other variants thereof are used in either the Detailed Description or the Claims, such terms are to be inclusive in a manner similar to the term "comprising." Further, it should be understood that aspects of the disclosure may be implemented in discrete circuits, partially integrated circuits, or fully integrated circuits or programming means. In addition, the terms "exemplary", "for example" and "z. B. "meant only as an example and not as best or optimal.

While specific aspects have been illustrated and described herein, it will be appreciated by those of ordinary skill in the art that various alternative and / or equivalent implementations may be substituted for the specific aspects shown and described without departing from the concept of the present disclosure. The present application is intended to cover any adaptations or variations of the specific aspects discussed herein.

Although the elements in the following claims are given in a particular sequence with corresponding designation, these elements, unless the claims otherwise require a particular sequence to implement any or all of these elements, are not necessarily limited to being implemented in that particular sequence become.

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited non-patent literature

• Baoguo Yang; Letaief, KB; Cheng, RS; Zhigang Cao: Channel estimation for OFDM transmission in multi-path fading channels based on parametric channel modeling, Communications, IEEE Transactions on, Vol. 49, No. 3, pp. 467-479, March 2001 [0035]

Claims (20)

Procedure ( 700 ) for estimating multipath channel parameters, the method ( 700 ) Includes: receiving ( 701 ) a signal comprising reference signal transmissions via a multipath channel; Provide ( 702 ) of the received signal in a first time domain representation; Detect ( 703 ) a first set of signal peaks based on the first time domain representation of the received signal; Provide ( 704 ) a pool of multipath channel parameters based on the detected first set of signal peaks; Choose ( 705 ) a certain amount of multipath channel parameters from the pool of multipath channel parameters according to an optimality criterion. Procedure ( 700 ) according to claim 1, wherein said providing ( 702 ) of the received signal in the first time domain representation, the application of an inverse discrete Fourier transform ( 321 ) to the received signal. Procedure ( 700 ) according to claim 1 or 2, wherein the reference signal transmissions comprise orthogonal frequency division multiplexed transmissions of a cell specific reference signal. Procedure ( 700 ) according to any one of the preceding claims, wherein said detecting ( 703 ) of the first set of signal peaks is based on adaptive threshold crossing of an average power spectrum of the first time domain representation of the received signal. Procedure ( 700 ) according to any one of the preceding claims, wherein the providing ( 704 ) of the pool of multipath channel parameters comprises: providing the received signal in a second time domain representation ( 325 wherein a granularity of the received signal in the second time domain representation in a range of the signal peaks of the first set of signal peaks is higher than a granularity of the received signal in the first time domain representation; Detecting a second set of signal peaks based on threshold crossing of an average power spectrum of the second time domain representation of the received signal; and providing the pool of multipath channel parameters based on the detected second set of signal peaks. Procedure ( 700 ) according to one of the preceding claims, wherein the pool of multipath channel parameters comprises channel taps (L) and respective delay values (τ) associated with the channel taps (L). Procedure ( 700 ) according to claim 6, wherein said selecting ( 705 ) of the determined set of multipath channel parameters comprises: starting from an initial number of channel taps and increasing the initial number in an iterative manner until a stop criterion is met. Procedure ( 700 ) according to claim 6 or 7, wherein said selecting ( 705 ) the determined set of multipath channel parameters comprises: selectively adding and removing channel taps of the pool of multipath channel parameters to the determined set of multipath channel parameters for each iteration. Procedure ( 700 ) according to one of the preceding claims, wherein the optimality criterion is based on an extended noise reduction algorithm. Procedure ( 800 ) for multipath channel parameter estimation, the method comprising: receiving ( 801 ) a signal comprising at least one demodulated cell-specific reference signal; Provide ( 802 ) a first and a second averaged power spectrum of a time domain transform of the received signal, wherein a granularity of the second average power spectrum in a range of signal peaks of the first averaged power spectrum is higher than a granularity of the first averaged power spectrum; Provide ( 803 ) a pool of multipath channel parameters comprising channel taps and associated delay values based on peaks of the second average power spectrum; and Determine ( 804 ) a set of multipath channel parameters from the pool of multipath channel parameters based on an optimality criterion. Procedure ( 800 ) according to claim 10, comprising: determining candidate sets of multipath channel parameters based on forward and backward selection ( 500 . 342 ) through the pool of multipath channel parameters. Procedure ( 800 ) according to claim 11, comprising: performing the forward and backward selection ( 500 . 342 ) from an initial number of channel taps in an iterative manner. Procedure ( 800 ) according to claim 12, comprising: selecting combinations of channel taps from the pool of multipath channel parameters according to a current iteration. Procedure ( 800 ) according to one of claims 11 to 13, comprising: determining mean square errors ( 505 ) between the at least one demodulated cell-specific reference signal and path gain estimates of respective candidate sets of multipath channel parameters. Procedure ( 800 ) according to claim 14, comprising: performing the forward and backward selection ( 500 . 342 ) through the pool of multipath channel parameters such that the mean square errors decrease. Procedure ( 800 ) according to claim 14 or 15, comprising: interrupting the forward and backward selection when a decrease in the mean square errors exceeds a threshold. Procedure ( 800 ) according to any one of claims 14 to 16, comprising: determining the path gain estimates by using parametric channel estimation based on a respective candidate set of multipath channel parameters. Procedure ( 800 ) according to one of claims 14 to 17, comprising: determining the path gain estimates by using the following parameters: a delay transformation matrix (T) representing propagation delays for the channel taps of a respective candidate set of multipath channel parameters; an autocovariance matrix (P) of gain factors of the channel taps of the respective candidate set of multipath channel parameters; and a noise power (σ 2 / w) the at least one demodulated cell-specific reference signal. Channel estimation device ( 900 ), comprising: a first unit ( 901 ) adapted to receive a signal ( 902 ) comprising at least one demodulated cell-specific reference signal; a second unit ( 903 ) designed to provide a pool of multipath channel parameters ( 906 ) based on the received signal ( 904 ); and a third unit ( 905 ) designed to determine a set of multipath channel parameters ( 908 ) from the pool of multipath channel parameters ( 906 ) on the basis of an optimality criterion. Channel estimation device ( 900 ) according to claim 19, wherein the third unit ( 905 ) is designed to determine the set of multipath channel parameters ( 908 ) based on a mean squared error between the at least one demodulated cell specific reference signal and a result of a candidate set of multipath channel parameters from the pool of multipath channel parameters ( 906 ) applied extended noise reduction algorithm is calculated.

Images