HK1164582B - Channel quality determination of a wireless communication channel based on received data - Google Patents

Description

Channel quality determination for wireless communication channels based on received data

Technical Field

The present invention relates to channel characteristic determination, and in particular to channel quality estimation of a wireless communication channel within a mobile communication network.

Background

In a wireless communication system, a change in the signal strength of a communication channel (due to multipath propagation changes, or due to shadowing from obstacles) may occur, for example, due to a change in the environment caused by movement of the wireless terminal; such channels are also known as fading channels. The channel quality of a fading channel may vary over time, frequency, and space. A transmitter, e.g. a base station of a mobile communication network, can use it to optimize data transmission if the channel quality is accurately estimated at the receiver side, e.g. one of the terminals. In particular, in a Frequency Division Duplex (FDD) system including an orthogonal frequency division multiplexing, OFDM, based system, a terminal may estimate channel quality to be fed back to a transmitter within a reasonably short feedback time. If the transmitter has knowledge of the downlink channel quality, the average throughput (and hence spectral efficiency) at the receiver side can be maximized while maintaining certain quality of service (QoS) parameters, such as guaranteed bit error rate.

A general problem with channel quality estimation is to estimate the block error rate (BLER) for data packets transmitted over a communication channel using a plurality of sub-channels, in particular in OFDM systems, wherein the communication channel is divided into a plurality of (narrow-band) sub-carriers, which allows orthogonal modulated data streams to be transmitted in parallel on the sub-carriers, provided that the current propagation channel conditions in a frequency selective fading channel have a different signal-to-interference-and-noise ratio (SINR) per sub-carrier.

There are well-known methods for CQI estimation discussed in the literature, for example, the method known as Exponential Effective SNR Mapping (EESM) described in the document entitled "System-level evaluation of OFDM-user connectivity" published by 3GPP under document number TSG-RAN WG1, R1-031303 (11 months, 17-21, 2003), and the method known as Mutual Information Effective SNR Mapping (MIESM) described in the document entitled "Effective-SNR mapping for model frame error rates in multiple-stations" published by 3GPP under document number C30-20030429-010, WG RAN1 (2003). Both methods use a reference signal, i.e. the estimated channel and noise variance is used to calculate the channel quality indicator CQI.

If a linear receiver (e.g. zero forcing or minimum mean square error) is employed, the effective SNR value is calculated by post-processing SINR and post-processing based SINR, e.g. estimated effective channel and noise variance based on (cell-specific) reference signals may be used for CQI determination.

However, CQI estimation based on a common (cell-specific) reference signal may not be accurate enough if the channel estimation error is not accurately accounted for. Furthermore, if a maximum likelihood detector is used instead of a linear equalizer, the CQI cannot be estimated using the EESM or MIESM methods because such detectors cannot deliver SINR.

Disclosure of Invention

It is an object of the invention to improve the determination of the channel quality.

This object is achieved by the independent claims. Advantageous embodiments are described in the dependent claims.

In an embodiment, the channel quality of a communication channel is determined at the receiver side by demapping a sequence of received modulation symbols encoded with a plurality of information bits to a plurality of corresponding soft values, calculating an (overall) mutual information MI metric as a function of the plurality of soft values, wherein the MI metric indicates interdependencies between the plurality of information bits and the plurality of corresponding soft values, also referred to as soft bits, which represent reliability information relative to a corresponding "hard decision" of whether an information bit is a "1" or a "0". Further, a channel quality indication, CQI, value is determined based on the MI metric.

Mutual information is a term that is commonly used in information or probability theory for the measure of interdependence of two (random) variables. The (overall) mutual information as described above will indicate the interdependence with respect to the plurality of soft bits. Thus, the (overall) mutual information may be a function of a plurality of specific mutual information values for each of the plurality of information bits. In an embodiment, the (overall) MI value is an average of the plurality of specific MI values. Alternatively, other functions, such as the square root of the sum of squared specific MI values, may be selected to form the (overall) MI value.

CQI estimation schemes are generally applicable to communication systems in which channel quality needs to be estimated. The above-described embodiments allow the channel quality CQI value to be derived based on the actual data being received. In case of beamforming with dedicated reference signals (receiver specific beamforming), the CQI of the receiver specific propagation channel will be estimated taking into account the actual data transmission.

Unlike CQI estimation based on user-specific reference signals, the proposed scheme is continuously applicable and may therefore yield more accurate results, especially if non-linear receivers are employed in the system (note that CQI estimation based on user-specific reference signals can only be evaluated during actual data transmission and can therefore only be based on a few dedicated reference symbols).

The MI metric may be determined over all information bits of one resource or transport block. In this case, the MI metric is a value indicating the interdependence of the entire transport block.

In OFDM-based multiple access systems with frequency selective channels, CQI is often used per subband (or per bandwidth part) for frequency selective scheduling of users. Thus, in an embodiment, if the full bandwidth is allocated to the user, the subband-specific CQI can be calculated by ordering the soft values according to their subbands and processing the subband-specific soft values separately.

In an embodiment, the soft values are log-likelihood ratios (i.e., log-binary of likelihood ratios), also referred to as log-likelihood ratios (LLRs), preferably computed by demapping the received modulation symbols.

In further embodiments, determining the channel quality indicator value is performed by comparing the MI metric to a plurality of particular threshold values, and determining a maximum threshold value out of the particular threshold values that is lower than (or equal to) the MI metric to be selected as the channel quality indicator value and reported to the wireless transmitter. The threshold may be chosen to yield a defined transport block error probability, e.g., a BLER below 10%.

In further embodiments, different sets of thresholds are selected for different modulation schemes applied for data transmission; for example, a first set of thresholds is used for QPSK, a second set of thresholds is used for 16QAM, and a third set of thresholds is used for 64 QAM.

In a further embodiment, the MI measure is determined by calculating an average over a plurality of particular MI measures each calculated as a function of the absolute value of one of the soft values, wherein the soft values are preferably so-called log-likelihood ratios each associated with a corresponding one of the information bits encoded in the data stream.

In further embodiments, the threshold value depends on at least one of: actual modulation and coding scheme, actual transport block size, specific power setting for actual transmission, and hypothetical power setting for determining CQI.

In further embodiments, the communication channel comprises a plurality of sub-channels with possibly different sub-channel propagation characteristics, e.g. exhibiting different signal to interference and noise ratios (SINRs). The subchannels may be implemented as subcarriers on which a plurality of orthogonally modulated data streams are transmitted in parallel according to OFDM.

In a further embodiment, in a multiple-input multiple-output, MIMO, transmission environment with a plurality of different codewords, a separate MI metric and corresponding CQI value to be fed back to the transmitter (1) is determined for each codeword.

In one embodiment, the proposed metric reuses the log-likelihood ratios required for data decoding.

Nearby CQIs can be introduced to indicate whether transmission with dedicated reference signals is preferable to transmission with common reference signals.

The invention also relates to a computer program comprising software code portions for implementing the method as described above when operated by respective processing units of a user device and a receiving device. The computer program can be stored on a computer readable medium. The computer readable medium can be a persistent or rewritable memory within or external to the user device or receiving device. The respective computer program can also be transmitted to the user device or the receiver device, for example as a signal sequence, via a cable or a wireless link.

Detailed embodiments of the present invention will be described below in order to provide a full and complete understanding to the skilled person. However, these examples are illustrative and not intended to be limiting.

Brief Description of Drawings

Figure 1 shows a schematic diagram of a feedback mechanism between a receiver and a transmitter of a communication system,

fig. 2 shows an exemplary block diagram of a receiver, which includes processing circuitry for determining a CQI-index (CQI-index) associated with a property of a transmission channel to be fed back to a transmitter,

figure 3 illustrates one embodiment of determining a differential CQI,

fig. 4 shows an exemplary sequence diagram of the sequence performed in the receiver according to fig. 2.

Detailed Description

Fig. 1 illustrates a schematic block diagram for illustrating the concept of channel quality feedback within a mobile communication network according to an embodiment of the present invention. Therein, fig. 1 shows a transmitter 1, a communication channel 2 and a receiver 3. By way of example, the modulation mapping circuit 11 is shown as being comprised by the transmitter 1, and the modulation demapping circuit 31 and the quality feedback circuit 32 and the data decoder 33 are shown as being comprised by the receiver 3. As an example, the modulation mapping circuit 11 receives the information bit sequence x₁，...，x_nAnd encodes them into a sequence s of modulation symbols to be transmitted over the communication channel 2 to the receiver 3₁，...，s_m. The modulation demapping circuit 31 receives a signal corresponding to a modulation symbol s₁，...，s_mOf the received value y₁，...，y_m(but they are typically different due to channel characteristics and noise added to the transmitted signal). It is noted that depending on the modulation scheme, a certain number of information bits are mapped onto one modulation symbol, e.g. 2 information bits are mapped into one QPSK symbol, 4 bits are mapped into one 16QAM symbol, and 6 bits are mapped into one 64QAM symbol. The general task of the receiver 3 is to encode the information bits x₁，...，x_nAnd (6) decoding.

Modern radio access systems, such as OFDM systems as described above, employ some forward error correction schemes, such as convolutional codes and convolutional turbo (turbo) codes. By using reliability information instead of performing hard decisions, receiver performance can be improved. The hard decisions and their reliability values are usually represented by a single so-called soft value, also called soft bit. For example, soft bits are used as input to turbo decoding based on Maximum A Posteriori (MAP) and Maximum Likelihood (ML) decoding rules, respectively. Given a modulation method such as, for example, Quadrature Phase Shift Keying (QPSK) or Quadrature Amplitude Modulation (QAM), log-likelihood ratios (LLRs) are used as soft bits to be provided to the decoder. According to such embodiments, the modulation demapping circuit 31 will receive the value y₁，...，y_mDemapping into a demapped sequence of soft values LLR₁，...，LLR_n. The data decoder 33 receives the demapped soft value LLR₁，...，LLR_nTo recover the information bit sequence x₁，...，x_n(i.e., the sequence of most likely information bits is generated or estimated).

As discussed in the introduction, the channel quality of a transmission channel may vary over time, frequency and space. The transmitter can use it to optimize the data transmission if the channel quality is estimated accurately at the receiver side. Accordingly, in practical mobile communication systems (e.g. LTE based), information about the actual channel quality is typically obtained by mobile terminals that periodically generate so-called Channel Quality Indicators (CQIs) that are fed back to the base station. It is to be noted that the CQI does not necessarily explicitly indicate the channel quality, but the data rate supported by the receiver under the current channel conditions. Accordingly, within the scope of the present application, the term CQI or CQI index should be interpreted broadly as any value (e.g. data rate, modulation scheme, transport block size, etc.) based on the measured channel conditions to be fed back from the receiver to the transmitter in order to set or adjust the data transmission. Specifically, the CQI may be information that satisfies a certain channel condition. Such information may be encoded in a number of bits (e.g., 5 bits) representing one CQI index out of a number of predetermined indices.

With the aid of the following fig. 2, the CQI determination within the receiver will be described in more detail. Typically, in an APP (a posteriori probability) processor, the soft input and soft output are a posteriori probabilities of the encoded information. By means of a so-called LogAPP algorithm (also called LogMAP) working directly on log-likelihood ratios (LLRs), the APP decoding can be in the log domain. The output of the LogAPP algorithm is the a posteriori LLRs of the information bits. Further Information on LogAPP decoding can be derived from the document entitled "calculation of Symbol-Wise Multi Information in Transmission Systems with LogAPP Decoders and Application to EXIT channels" (Proc.5th int. ITG Conf. on Source and Channel Coding (SCC), Erlangen-Numberg, Germany, 1 month 2004, pp.195-202) of lngmar Land, Peter A. Hoeher and Snjezana Gligravic.

Fig. 2 shows the modulation demapping circuit 31 of fig. 1, a CQI feedback circuit exemplarily distributed into the mutual information MI determining circuit 32a and the CQI index determining circuit 32b, and the threshold setting circuit 33.

The symbol demapping circuit 31 receives the transmitted symbol y₁，...，y_m. Further, this circuit receives modulation information MD indicating actual modulation for actual transmission, for example, QPSK, 16QAM, 64QAM, or the like. To obtain the log-likelihood ratio LLR (k), the modulation symbol y₁，...，y_mAre demapped to so-called soft bits. As an example, the demapping circuit 31 is a so-called LogMAP/LogAPP demapper, presenting at the output the log-likelihood ratios llr (k) as soft bits.

The log-likelihood ratio may further be used as input to a data decoder such as shown in fig. 1, e.g., a possible rate dematching and HARQ combiner, followed by a turbo decoder (e.g., operating on LogMAP metrics).

The MI determining circuit 32a determines overall statistical information or overall mutual information MI from a set of k ═ N specific mutual information values MI (1),. -, MI (N) obtained from the corresponding log-likelihood ratios.

From the absolute value of the kth log-likelihood ratio, i.e., | llr (k) |, of the reliability information indicating the kth log-likelihood ratio, the kth mutual information mi (k) can be extracted:

where Id will represent a logarithmic binary.

It is to be noted that although the kth specific mutual information mi (k) is obtained from the transmitted actual data, it does not depend on the actual values of the information bits.

To derive the overall metric statistics, the overall MI value is obtained from MI (k), e.g. by averaging MI (k) over so-called CQI reference resources, which are resources in time and frequency where CQI is to be estimated:

the MI thus obtained includes potential impairments from a defective receiver front end and channel estimation as well as the characteristics of the physical propagation channel.

The CQI index determination circuit 32b receives the mutual information MI and a certain number of threshold sets TH QPSK, TH16QAM, and TH64QAM supplied from the threshold setting circuit 33.

As an example, the CQI index determination circuit 32b includes a first comparison circuit 321 for comparing MI with a first threshold set TH QPSK, a second comparison circuit 322 for comparing MI with a second threshold set TH16QAM, and a third comparison circuit 323 for comparing MI with a third threshold set TH64 QAM. Depending on the modulation considered, a corresponding one of these circuits is activated to determine the CQI index.

The CQI index to be reported is calculated by a simple thresholding and maximum value operation as an example:

in other words, the CQI index to be reported is the largest threshold value smaller than MI out of all threshold values of the corresponding set. If, for example, QPSK is used, the first comparison circuit 321 is activated to determine a threshold ldx from a first set of thresholds TH QPSK₁，...，ldx_x1Maximum threshold value Idx of_i。

It is noted that each threshold corresponds to a particular channel condition and represents a particular CQI index that can be reported. The threshold may correspond to a minimum MI where a transmission with a certain BLER, e.g. 10%, is successful under a given set of conditions for the transmission.

Note also that the CQI may not necessarily indicate the channel quality explicitly, but rather the data rate supported by the receiver under the current channel conditions. More specifically, the CQI may be a suggested transport block size (corresponding to a suggested data rate). According to the current LTE specifications, CQI is a 5-bit value to be periodically fed back from a user equipment (receiver) to a NodeB (transmitter).

Averaging the mi (k) values yields a suitable metric for CQI estimation, while averaging the corresponding received SINR on the CQI reference resource will not yield a useful metric for CQI estimation due to the high dynamics of the received SINR across the subcarriers.

The threshold setting circuit 33 may determine the threshold offline, for example by means of a calculation by simulation. This circuit may be further adjusted on-line, for example, using long-term statistics such as BLER.

The threshold may depend on the following parameters:

the modulation scheme MD employed in the actual transmission. As discussed above, a different set of thresholds may be used for each modulation scheme used in the actual transmission,

the modulation and coding scheme defined for a particular CQI index. It is noted that the transport block size may also vary depending on the modulation and coding scheme selected.

The actual power setting PA and the reference power setting PR used for CQI estimation, or e.g. the offset between the power in the actual data transmission and the reference power. The reference power for CQI estimation is defined in some standards, while the power in the actual transmission (or the power offset between the reference signal and the data signal) is typically known in the receiver. Communication standards, such as 3GPP documents, often specify a set of predetermined transmission powers and therefore only a separate set of power offsets need to be considered.

In the following, exemplary extensions to per-codeword CQI computation and subband CQI computation are described:

in OFDM-based multiple access systems with frequency selective channels, CQI is often used per sub-band (per bandwidth part) for frequency selective scheduling of users. If the full bandwidth is allocated to the user, a subband-specific CQI can be calculated by ordering the LLR values according to their subbands and processing the subband-specific LLR values separately (as shown in the previous section).

In MIMO (multiple input and multiple output) transmission with more than one codeword, CQI estimation can be performed for each codeword separately. In particular, the MI value for each codeword is processed separately as described above to generate two separate CQIs that are fed back to the transmitter.

To decide between transmissions with dedicated reference signals (also called beamforming) or common reference signals, differential CQI values indicating the gain or loss from data throughput using dedicated reference signals versus data throughput using common reference signals are proposed. This decision can be used to adaptively switch between transmissions with a common reference signal and transmissions with a dedicated reference signal if the communication system allows switching between transmissions with a common reference signal and transmissions with a dedicated reference signal, as in LTE.

The CQI for transmission with a common reference signal can be obtained by well-known techniques, e.g. using the EESM or MIESM methods as described in the background section. In particular, the common reference signal is continuously transmitted and thus the CQI for the common reference signal can always be evaluated in parallel with the CQI for the transmission with the dedicated reference signal. Note that the data throughput (and hence CQI) for transmission with dedicated reference signals may be higher (due to improved channel conditions due to user-specific beamforming) or lower (due to increased overhead due to the use of dedicated reference signals).

In an embodiment, instead of transmitting separate CQI values for each of the transmission with the common reference signal and the transmission with the dedicated reference signal, only one CQI value is transmitted along with the differential CQI value. Fig. 3 shows an exemplary block diagram with a first CQI determination circuit 41, a second CQI determination circuit 42, and a difference circuit (difference circuit) 43. The first CQI determination circuit 41 determines a first CQI value CQI index 1 based on actually received data. To this extent, this circuit may employ the functionality described in accordance with fig. 3, in particular the functionality of the MI determining circuit 32a, the CQI index determining circuit 32b and the threshold setting circuit 33. The second CQI determination circuit 42 includes, as an example, a CQI estimation circuit 421 and a thresholding circuit 422. The CQI estimation circuit 421 estimates a CQI value based on a common reference signal, for example, using the EESM or MIESM method as described above. The estimated CQI is fed to a thresholding circuit 422 that selects a maximum threshold value below the estimated CQI to determine a second CQI value CQI index 2. The first CQI value CQI index 1 and the second CQI value CQI index 2 are each fed to a differential circuit 43 which determines a differential CQI index which is reported, as an example, together with the second CQI value CQI index 2 based on a common reference signal.

This embodiment allows to reduce the feedback data rate. The differential CQI may thus include the case where only a 1-bit flag is transmitted, indicating whether transmission with a dedicated or common reference signal is preferred.

The method has particular advantages for the case of beamforming with dedicated (UE-specific) reference signals, where the CQI of the beamformed propagation channel needs to be calculated based on the actual data transmission.

Fig. 4 shows a flow chart with exemplary outlined steps for determining CQI in a receiver as described according to fig. 1:

in a first step 41, a log-likelihood ratio llr (k) is determined from the received modulation symbols.

In a second step 42, a specific mutual information value mi (k) is calculated for each llr (k).

In a third step 43, the (overall) mutual information MI is calculated as a summary metric of the specific mutual information values MI (k), e.g. by averaging the specific mutual information values MI (k).

In a fourth step 44 the derived (overall MI) value is compared to a plurality of thresholds, e.g. each threshold representing a certain channel condition.

In a fifth step 45, a maximum threshold value below the derived MI value is determined as the CQI index to be reported to the transmitter.

Claims (11)

1. A method of determining a channel quality of a communication channel (2) between a wireless transmitter (1) and a wireless receiver (3), comprising:

-receiving a signal with one or more modulation symbols (y)₁,...,y_m) The modulation symbols comprise a plurality of coded information bits (x)₁,...,x_n)，

-demapping the modulation symbols to a plurality of soft values (LLRs)₁,...,LLR_n)，

-as the plurality of soft values (LLR)₁,...,LLR_n) Is used to calculate a mutual information MI metric, wherein the MI metric indicates information bits (x) of the transport block₁,...,x_n) Interdependence with corresponding soft values (LLR (k)), and

-determining a channel quality indicator value, CQI, index 1 as a function of the MI metric,

wherein the channel quality indicator value, CQI, index is determined by comparing the MI metric with a plurality of specific thresholds TH QPSK, TH16QAM, TH64QAM and determining a maximum threshold out of the specific thresholds below the MI metric,

wherein the determined maximum threshold value is selected as a channel quality indicator value, CQI, index to be reported to the wireless transmitter (1).

2. The method of claim 1, wherein the particular threshold is selected to yield a transport block error probability of no more than 10%.

3. The method of claim 1, wherein the MI metric is determined by calculating an average over a plurality of particular MI metrics, each particular MI metric calculated as a function of an absolute value of one of the soft values.

4. The method of claim 1, wherein the soft values are log-likelihood ratios (LLRs)₁,...,LLR_n) The log-likelihood ratios are each associated with the information bit (x)₁,...,x_n) Is associated with a corresponding one of the bits.

5. The method of claim 1, wherein the communication channel comprises a plurality of subchannels with different subchannel propagation characteristics.

6. The method of claim 5, wherein the sub-channels are implemented as subcarriers, wherein if the communication channel exhibits selective fading, multiple orthogonally modulated data streams are transmitted in parallel on the subcarriers, each subcarrier exhibiting a different signal-to-interference-and-noise ratio (SINR).

7. The method of claim 1, wherein at least one of the specific thresholds depends on at least one of:

the modulation and coding scheme of the modulation and coding scheme,

the size of the transport block is then used,

a specific power setting for the actual transmission, and

a hypothetical power setting for determining the CQI.

8. The method of claim 1, wherein performing multiple-input and multiple-output MIMO transmission with a plurality of different codewords comprises determining a separate MI metric for each codeword separately and determining corresponding individual channel quality indicator values to be fed back to the wireless transmitter (1).

9. The method according to claim 1, wherein a further channel quality indicator value CQI index 2 is determined on the basis of a common reference signal, and wherein one of the channel quality indicator value CQI index 1 and the further channel quality indicator value CQI index 2 is transmitted together with a differential CQI value indicating a difference of the channel quality indicator value CQI index 1 and the further channel quality indicator value CQI index 2.

10. A wireless receiver (3) adapted to determine a channel quality of a communication channel (2) between a wireless transmitter (1) and the wireless receiver, comprising:

-a modulation demapping circuit (31) adapted to receive a signal having one or more modulation symbols (y)₁,...,y_m) The modulation symbols comprise a plurality of coded information bits (x)₁,...,x_n)，

-demapping the modulation symbols to a plurality of soft values (LLRs)₁,...,LLR_n)，

-a mutual information determination circuit (32a) adapted to determine the soft values (LLR) as the plurality of soft values₁,...,LLR_n) Is used to calculate a mutual information MI metric, wherein the MI metric indicates information bits (x) of the transport block₁,...,x_n) Interdependence with corresponding soft values (LLR (k)), and

-a CQI index determination circuit (32b) adapted to determine a channel quality indicator value CQI index as a function of the MI metric,

wherein the channel quality indicator value, CQI, index is determined by comparing the MI metric with a plurality of specific thresholds TH QPSK, TH16QAM, TH64QAM and determining a maximum threshold out of the specific thresholds below the MI metric,

wherein the determined maximum threshold value is selected as a channel quality indicator value, CQI, index to be reported to the wireless transmitter (1).

11. An apparatus for determining a channel quality of a communication channel (2) between a wireless transmitter (1) and a wireless receiver (3), comprising:

-for receiving a signal with one or more modulation symbols (y)₁,...,y_m) The modulation symbols comprise a plurality of coded information bits (x)₁,...,x_n)，

-means for demapping the modulation symbols to a plurality of soft values (LLRs)₁,...,LLR_n) The apparatus of (1) is provided with a plurality of the devices,

-for use as the plurality of soft values (LLRs)₁,...,LLR_n) Means for calculating a mutual information MI metric, wherein the MI metric indicates information bits (x) of the transport block₁,...,x_n) Interdependence with corresponding soft values (LLR (k)), and

-means for determining a channel quality indicator value, CQI, index 1 as a function of the MI metric,

wherein the channel quality indicator value, CQI, index is determined by comparing the MI metric with a plurality of specific thresholds TH QPSK, TH16QAM, TH64QAM and determining a maximum threshold out of the specific thresholds below the MI metric,

wherein the determined maximum threshold value is selected as a channel quality indicator value, CQI, index to be reported to the wireless transmitter (1).