JP2002515707A - Channel estimation method and apparatus - Google Patents

Abstract

(57) Summary According to the present invention, in a channel estimation method in a communication system, a signal including a data portion having a data symbol and a midamble having a known symbol is transmitted via a transmission channel. At the receiving side, the channel coefficients are estimated with respect to the channel impulse response of the transmission channel, where the received signal that depends on the midamble is estimated from known symbols of the midamble to estimate the channel coefficients. Cyclic convolution with Fourier transform using coefficients. The Fourier transform is performed with an arbitrary number S of reference points, as long as S ≧ 2L, where L is the length of the evaluable part of the midamble. Particularly advantageously, the invention is used in third generation mobile radio networks.

Description

DETAILED DESCRIPTION OF THE INVENTION

The present invention relates to a communication system, for example, a channel estimation method and apparatus in a mobile radio network.

[0002] In communication systems, communications (eg, language, image information or other data) are transmitted over transmission channels, and in wireless communication systems this is done over a wireless interface using electromagnetic waves. The radiation of the electromagnetic waves takes place here at carrier frequencies in the frequency band set for the respective system. GSM (Global System for Mobile Communi
In cation), the carrier frequency is in the region of 900 MHz. Future wireless communication systems such as UMTS (Universal Mobile Telecommunication System)
For other systems of the third generation, the frequency is provided in a frequency band of about 2000 MHz.

The emitted electromagnetic waves are attenuated based on reflection, diffraction and radiation losses due to the curvature of the earth and the like. As a result, the reception power that can be used in the receiving radio station decreases. This attenuation is location dependent and also time dependent as the radio station moves. In the case of multipath propagation, a plurality of signal components arrive at a receiving wireless station with different delays. Such effects explain the connection-specific transmission channel.

[0004] From DE 195 49 158 a radio communication system using code division multiple access (CDMA) is known, wherein the radio interface is additionally time division multiplexed. (= Time Division Multiple Access, TDMA = Time Division Multiple Acces
s). In general, in the code multiplexing method, a code sequence (for example, in a form of superposition with a noise signal of a predetermined energy) is added to an individual signal assigned to an individual subscriber, and data to be transmitted by the individual subscriber is transmitted. They are separated. On the other hand, in time division multiplexing (TDMA), time slots that are transmitted in time sequence are allocated to various subscribers, and these time slots are grouped into frames. The time slot sequence is repeated. Further, at the receiving end, a joint detection (JD) method is used to enable improved detection of the transmitted data based on the individual subscriber's knowledge of the CDMA code. In the JD method, CDMA individual signals are commonly captured and detected and supplied to a matched filter. This filter is tuned to the respective individual signal or CDMA code of the individual subscriber, and the output signal of the matched filter is subsequently processed by maximum likelihood decoding to obtain the most likely output signal vector. Can be ascertained. Therefore, the JD method can remove an obstacle of an individual signal due to another individual signal.

[0005] Furthermore, the above-mentioned publication discloses that a connection can be assigned at least two data channels via a radio interface, wherein each data channel is distinguished by a unique spreading code. Can be.

In particular, when a mobile receiver, such as a mobile telephone, moves, in particular a superposition of different propagation paths of the subscriber signal occurs on the receiving side, whereby the transmission is carried out via different propagation paths. Individual signals of the same subscriber can be excluded even if the difference in travel delay time is excluded.
Since they are usually affected by different attenuations and distortions,
Due to the interference, the desired subscriber signal cannot be correctly reproduced at the receiving side from the received individual signals of the corresponding subscriber.

It is known from GSM mobile radio networks to transmit data to be transmitted as radio blocks (bursts) in time slots, in which a midamble with known symbols is transmitted in the radio block. These midambles can be used for tuning of the receiving side of the radio station in the sense of a training sequence. The receiving wireless station performs channel impulse response estimation for various transmission channels based on the midamble so that the receiving ability of the wireless station can be increased. In this case, the length of the midamble is fixedly determined irrespective of the traffic conditions.

If information of a plurality of connections is transmitted in one time slot at the same time as in the case of the TD / CDMA transmission method, channel estimation must be performed simultaneously for various transmission channels. In that case, the number of connections, and thus the number of channel impulse responses to be estimated, may fluctuate. As the number of connections or the length of the channel impulse response to be estimated changes, so does the range or spread of the estimation method, so that a large number of estimation methods must be implemented for all possible variations. Become.

It is therefore an object of the present invention to provide a channel estimation method and apparatus which allows an economical implementation even if the demands on the channel estimation change, with the channel estimation being particularly low. Want to be able to implement at cost.

This object is achieved by a method having the features of claim 1 and an apparatus having the features of claim 11. Advantageous embodiments of the invention can be taken from the dependent claims.

According to the present invention, in a channel estimation method in a communication system, a signal including a data portion having a data symbol and a midamble having a known symbol is transmitted via a transmission channel. At the receiving end, the channel coefficients are estimated with respect to the channel impulse response of the transmission channel, where the received signal, which depends on the midamble, is derived from known symbols of the midamble to estimate the channel coefficients. It is convolved cyclically using the coefficients.

Cyclic convolution can be advantageously implemented in terms of cost by means of a fast Fourier transform (FFT) or a combination of a fast Fourier transform (FFT) and an inverse fast Fourier transform (IFFT) in the frequency domain. The cyclic convolution can in principle be carried out with an arbitrary length, ie with an arbitrary number S of reference points. S ≧ 2L reference points are already sufficient, where L is the length of the evaluable part of the midamble. In particular, it is not necessary to select the number S of reference points as a power of two.

The Fourier transform with S ≧ 2L reference points is economically implemented in digital signal processing means such as a digital signal processor or ASIC. S ≧ 2L
, The midamble length L with L _M known symbols
For all the demands in the channel estimation that occur in the case of (a), it is possible to carry out a cyclic convolution without changing the algorithm, i.e. with different numbers of subscribers and with different channel impulse responses to be estimated. The same algorithm can be used for length.

When the estimation of the channel coefficient is performed in common for a plurality of connections, further simplification is realized. Thereby, the Fourier transform is directed only to the hassle of a perfectly aligned channel estimate of the midamble transmitted simultaneously in the frequency channel.

To avoid having to perform matching, channel coefficients are also estimated for transmission channels not used by the connection. This allows the Fourier transform to remain unaffected by changes in the number of connections over the wireless interface.

According to another advantageous aspect of the invention, scaling of the estimated channel coefficients is performed. Since multiple data channels can be assigned to one connection and the energy states or ratios of the data symbols and the associated channel coefficients of one data channel should be the same for all connections. , The necessary compensation is performed by scaling. Scaling improves the data estimation (detection) following the channel estimation.

The mapping of the estimated channel coefficients to one or more data channels and connections with connection-specific midambles is preferably carried out on the basis of a table. The configuration data of the wireless interface is entered in this table.

The method of the present invention is applicable to a wide variety of transmission channels (wired or non-wired). Particularly advantageously, the channel estimation is improved if the communication system is a mobile radio network and the rapidly changing transmission channel is a radio channel of a radio interface.

If the length L of the midamble to be evaluated is dynamically matched to the total length L _{M of the} midamble, the number of connections in the time slot and the length of the channel impulse response to be estimated are: On average, the spectral efficiency of the radio interface is increased. Nevertheless, the channel estimation of the present invention remains feasible when the number of reference points S is adjusted to the maximum possible L.

Furthermore, it is within the framework of the invention to derive the midamble used in the time slot from a common midamble basic code. This allows the midamble to be generated in a particularly simple manner on the transmitting and receiving sides and the channel estimation to be
The midamble can be implemented in common for all connections derived from a common midamble basic code.

It is advantageous to assign a plurality of data channels to one connection, wherein a smaller number of midambles is used than the number of data channels. This reduces the cost of channel estimation. Additionally, the number of possible data channels per time slot is increased. This is because multiple data channels use the same midamble and the effect of limiting the capacity of the channel estimation has no effect on the data channel.

Embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 shows a block diagram of a mobile radio network, FIG. 2 schematically shows a frame structure of a radio interface, FIG. 3 schematically shows a configuration of a radio block, and FIG. FIG. 5 shows a block diagram of a receiver of a radio station, FIG. 6 shows a block diagram of a digital signal processing means, and FIG. 7 shows a sequence line of channel estimation. The figure is shown.

The configuration of the wireless communication system shown in FIG. 1 corresponds to the configuration of a known GSM mobile wireless network. In other words, this means that a plurality of mobile switching centers MS
Consists of C. These mobile switching centers are either networked together or have access to a fixed network PSTN. Furthermore, each of the mobile switching centers MSC is connected to at least one base station controller BSC. Each base station controller BSC enables a connection to at least one base station BS. This type of base station BS can establish a wireless connection to the mobile station MS via a wireless interface.

FIG. 1 illustrates three wireless connections for transmitting valid information ni and signaling information si between three mobile stations MS and one base station BS, in which case one mobile station is used. The station MS is assigned two data channels DK1 and DK2 and another mobile station MS has one data channel DK3 each.
, DK4. The operation and maintenance center OMC performs control and maintenance functions for the mobile radio network or parts thereof. Although the functionality of this structure is utilized by the wireless communication system of the present invention, it can be transferred to other wireless communication systems that can use the present invention.

The base station BS is connected to an antenna device consisting of, for example, three individual radiators. Each of the individual radiators radiates and is directed to one sector of the radio cell served by the base station BS. However, alternatively, a larger number of individual radiators (according to the adaptive antenna) can be used, in which case it is possible to use spatial subscriber separation according to the SDMA (Space Division Multiple Access) scheme. it can.

The base station BS enables the mobile station MS to use organization information regarding the location area (LA location area) and the radio cell (radio cell beacon). These organization information are emitted simultaneously through all the individual radiators of the antenna device.

The connection between the base station BS and the mobile station MS having the valid information ni and the signaling information si is performed by multipath propagation. This is caused, for example, by reflections in the building in addition to the direct propagation path. By the directional radiation of the predetermined individual radiator of the antenna device AE, a relatively large antenna gain is obtained as compared with the omnidirectional radiation. Q of connection
Is improved by directional radiation.

Assuming that the mobile station MS moves, multipath propagation, combined with another obstacle,
At the receiving mobile station MS, signal components of different propagation paths of the subscriber signal are superimposed in a time-dependent manner. Furthermore, different base stations BS
Is superimposed on the received signal rx in the frequency channel at the receiving location. The task of the receiving mobile station MS is to detect the data symbol d of the valid information ni, the signaling information si and the data of the organization information transmitted in the subscriber signal.

The frame structure of the radio interface is shown in FIG. TDMA
According to -Koponente (component), division into a plurality of time slots ts, for example, eight time slots ts1 to ts8, in a wide frequency range, for example, a bandwidth B = 1.6 MHz is set. Each time slot ts in frequency domain B forms one radio channel. Information of a plurality of connections is transmitted in a radio block in a radio channel provided for valid data transmission. FDMA (Frequency Division Multiple Acce)
According to ss) -Componente (component), a plurality of frequency domains B are allocated to the wireless communication system.

As shown in FIG. 3, these radio blocks for valid data transmission consist of a data part dt with data symbols d. In these data parts, parts having a known midamble m on the receiving side are inserted. Data d
Are spread by a connection-specific detailed structure and a spreading code (CDMA code), so that, for example, K data channels DK1, DK2, DK3,... DKK can be separated on the receiving side by these CDMA components. It is. A predetermined energy E is assigned to these data channels DK1, DK2, DK3,.

The spreading of the individual symbols of the data d with Q chips results in the transmission of Q sub-parts of the duration Tc within the symbol duration Ts. At that time Q
Each chip forms an individual CDMA code. The midamble m consists of L chips, which also have a duration Tc. Furthermore, a guard time guard for compensating for different signal transit delay times of connections of successive time slots ts is provided in the time slot ts.

In the broadband frequency domain B, successive time slots ts are configured according to a frame structure. That is, eight timeslots ts are combined into one frame, where a given time slot of the frame forms one radio channel for valid data transmission and is repeatedly used by a group of connections. For example, another frequency channel for the frequency or time synchronization of the mobile station MS is inserted in the multiframe at a predetermined point in time, rather than in each frame. The spacing between these frequency channels defines the capacity that the wireless communication system can use for it.

The parameters of the radio interface are, for example: radio block duration 577 μs number of chips per midamble m 243 guard time Tg 32 μs data symbols per data part N 33 symbol duration Ts 6.46 μs symbol Q Parameters per chip 14 Chip duration Tc 6/13 μs In the up direction (MS → BS) and in the down direction (BS → MS), parameters can be adjusted and set differently.

If the midamble length is dynamically matched to the number M of connections in the time slot and the length W of the channel impulse response to be estimated,
On average, the spectral efficiency of the radio interface is increased. In doing so, it must be taken into account that only a limited number of channel impulse responses can be commonly estimated per time slot ts. This restriction is that the midamble contains L evaluable chips, the channel impulse response has W coefficients for accurate channel estimation, and M is the number of connections per time slot. It arises from expressing. In this case, the number h of the channel impulse types that can be commonly estimated is limited by the inequality L ≧ M * W + W−1.

By utilizing one common midamble m for the plurality of data channels DK1 and DK2 of the connections V1, V2, V3, it is possible to transmit more data channels DK1 and DK2 in the time slot ts. it can. This will either increase the data rate per time slot ts or extend the estimable channel impulse response in this time slot ts (eg for complex land structures).

The transmitter or receiver shown in FIGS. 4 and 5 relates to a radio station which can be a base station BS or a mobile station MS. At the receiver, the inventive device for channel estimation is used. However, in FIG. 4 and FIG.
Only the signal processing for 1 is shown.

FIG. 4 shows the transmission path of the device in detail. It is shown in customary form for modeling and simulation of communication technology systems. That is, the dependencies between various functions and the system structure are illustrated here.

[0039]

[Outside 1]

Modulated subscriber-specific CDMA code c ^{(k) for} 4PSK modulation and data
, K = 1... K, in the submodule S3. Thereafter, the addition of all spread data sequences takes place in the submodule S4 and subsequently the integration of the midamble m into the burst structure in the submodule S5. In the submodule S6, the spectral formation of the transmission signal s follows. In the modules S7 to S9, the transmission signal which has been time-discretely oversampled four times in the baseband s is converted into a time- and value-continuous bandpass region of the transmission frequency band.

To illustrate the function of the transmitter, the transmitter receives a digitized data symbol d before it at the data source (microphone or network-side connection), in each case N = 33 data symbols d. Are separately processed. First, rate 1/2 and constraint length 5 channel coding is performed in a convolutional encoder, followed by scrambling in an interleaving device having a scrambling depth of 4 or 16.

The scrambled data is subsequently 4PSK modulated in a modulator, converted to 4PSK symbols and spread therefrom in spreading means corresponding to the individual CDMA codes. This processing is performed in parallel on all data channels DK1 and DK2 of the connection V1 in the signal processing means DSP. In the case of the base station BS, the other connections V2 and V3 are similarly processed in parallel. The digital signal processing means DSP includes a control device S as shown in FIG.
It can be realized by digital signal processors DSP1, DSP2, DSP3 controlled by E.

In the adder element, the spread data of the data channels DK1 and DK2 are superimposed, in which the data channels DK1 and DK2 are equally weighted. A time-discrete representation of the transmitted signal for the mth subscriber can be made according to the following equation:

[0044]

(Equation 1)

Where K (m) is the number of the mth subscriber's data channel and N is the number of data symbols d per data part dt. The superimposed subscriber signal is supplied to a radio block former (burst former). The radio block former organizes the radio blocks taking into account the midamble m specific to the connection.

Complex CDM derived from a binary CDMA code by multiplication with j ^{q -1}
Since the A code is used, the output signal of the Chip Impulse Filter following the radio block former is GMSK modulated and has a substantially constant envelope if the connection uses only one data channel. are doing. The chip impulse filter performs convolution with the GMSK main impulse.

Following the digital signal processing, D / A conversion, transmission in a transmission frequency band, and signal amplification are performed on the transmission side. Subsequently, the transmission signal tx reaches a radio station, for example a mobile station MS, which is emitted via an antenna arrangement and possibly received via different transmission channels.

The individual midamble m consisting of L complex chips is used per connection. The M different midambles required are derived from a basic midamble code of length M * W, where M is the maximum number of subscribers (connections) and W is the channel coefficient h of the channel impulse response. Represents the maximum number expected. The midamble m specific to the connection is the right rotation of the basic midamble code by W * m chips and L ≧ (M + 1) * W
Derived by periodic expansion up to -1 chip. Since the complex basic midamble code is derived from the binary midamble code by modulation with j ^q−1 , the transmitted signal of midamble m is similarly GMSK modulated.

FIG. 5 shows the receiving path of the device in detail. In the partial module E1, the conversion of the received signal rx from the transmission frequency band to the low-pass region and the division into real and imaginary components are performed. In the submodule E2, analog low-pass filtering is performed and in the submodule E3 a further 13/3
Two oversamplings of the received signal with MHz and 12 bit word width are performed.

In the partial module E 4, digital low-pass filtering is performed by a filter having a bandwidth of 13/6 MHz having the highest possible edge steepness for channel separation. Subsequently, a 2: 1 decimation of the signal twice oversampled is performed in the submodule E4.

The received signal e thus obtained consists essentially of two parts, a component em for channel estimation and components e1 and e2 for data estimation. In the partial module E5, all channel impulse responses h ^(k)
Is estimated using the known midamble base code m of all data channels transmitted in each time slot.

In the submodule E 6, the parameter b ^(k) for the matched filter
Are determined using the CDMA code c ^(k) for each data channel. In the sub-module E7, the cancellation of interference in the receive block e1 / 2 used for data estimation, which is derived from the midamble m ^(k), is performed. This is possible with knowledge of h ^(k) and m ^(k) .

In the submodule E 8, the calculation of the cross-correlation matrix A ^{* TA} is performed.
Since A ^{* TA} has a tapelit structure, only a small portion of the matrix needs to be calculated here. This part can then be used for expansion to a complete volume. In the submodule E9, a Cholesky decomposition of A ^{* T} A into H ^{* T} H is performed, where H is the upper triangular matrix. Based on the A ^{* T} A tapelitz structure, H also has approximately the tapelitz structure and need not be fully calculated. The vector s represents the reciprocal of the diagonal element of H. These can advantageously be used in a system of solving simultaneous equations.

In the sub-module E 10, matched filtering (matched filter) of the received symbol sequence e1 / 2 having b ^(k ) is performed. Partial module E
Reference numeral 11 denotes a simultaneous equation solving apparatus 1 for H ^{* T} * z1 / 2 = e1 / 2, and a partial module E12 denotes a simultaneous equation solving apparatus 2 for H * d1 / 2 = z1 / 2. . In the partial module E13, the estimated data d1
/ 2 is demodulated, descrambled and further convolutionally decoded using a Viterbi decoder. Decrypted data block

[0055]

(Equation 2)

Optionally, a second data sink D 1 or a second data sink D 1 via a source decoder E 14.
Is supplied to the data sink D2. Source decoding is performed on the signaling channel SA.
Required for data blocks transmitted over CCH or FACCH.

On the receiving side (see FIG. 5), after analog processing, ie amplification, filtering, and conversion to baseband in the high frequency section, digital filtering of the received signal rx is performed in a digital filter. A digitized received signal e represented by a vector em of length L = M * W and free of data part dt interference
Are transmitted to the channel estimator. The common channel estimate of all M channel impulse responses is detailed in FIG.

Data estimation in the joint detection data estimator is performed in common for all connections. The CDMA code is represented by the received data with c ^(k) , d ^(k) and the corresponding channel impulse response with h ^(k) , where k = 1 to K.

The part of the received signal used for data estimation is described by the vector e = A · d + n, where A is the CDMA code c ^(k) known a priori and the estimated channel impulse response h ^(k) is the system matrix. The vector d is a combination of the data d ^(k) for each data channel as shown in the following equation:

[0060]

(Equation 3)

For this symbol arrangement, system matrix A has a band structure used to reduce the complexity of the algorithm. The vector n is
Contains noise components. Data estimation is based on the following formula: Zero Forcing Block Linear Equalizer (ZF-BLE)
Implemented by: d = (A ^{* T} A) ^-1 A ^{* T} e.

These components are continuous values, not manipulated estimates of the data symbol d. To simplify the calculation of d, the problem can be rewritten as a system of linear equations of the form (A ^{* TA} ) d = A ^* Te, where Cholesky's decomposition A ^{* TA} = H ^{* According to T} H, the calculation of the data symbol d is reduced to a solution of the following two systems of linear equations, H ^{* T} z = A ^{* T} e, where H · d = z. The solution of this simultaneous equation can be implemented cyclically.
H is the upper triangular matrix and H ^{* T} is the lower triangular matrix.

FIG. 6 shows a digital signal processing means DSP. It receives the received signal rx already digitized at the receiving side and sends out the estimated data symbols d. The digital signal processing means DSP includes a plurality of digital signal processors DSP1, DSP2, DSP3, a memory SP, and a control device SE.

The memory SP stores a table T1 and an estimation coefficient g ″ ′ for channel estimation, which will be described later. One of the digital signal processors DSP2 controls the channel estimator KS by a corresponding program module. Has been realized.

The channel estimator KS transmits from the received sequence em of length L ≧ L _M −W + 1, which depends only on the midamble m using the channel coefficient h of length K of all K subscribers. Used to estimate the channel, where L _M = 243 and W = 27
, The length is equal to 217. L is the length of the part to be evaluated of the received signal em that depends on the midamble m. If L = L _M -W + 1, the energy of the midamble m is optimally utilized. For example, the length used is M = 8 * 2
7, which is equal to 216. Channel estimation with inverse filtering makes use of the knowledge about the midamble m ^(M) used by the mobile station MS, resulting from the cyclic midamble code m ^W. The used cyclic midamble code of length L depends on the number W of channel coefficients h to be estimated.

The mobile station MS in the connection V 1 has a plurality of data channels DK 1, DK 2,.
, The estimated channel impulse response is the data channel DK1
, DK2,..., DK2,..., DK2,. This is done by scaling with a scaling factor that depends on the total power of the channel impulse response.

As described later with reference to FIG. 7, S> W * required for channel estimation
If cyclic convolution by a discrete Fourier transform / inverse Fourier transform with K (FFT / IFFT) is to be performed, it is sufficient to select S ≧ 2L. In particular, S need not be a power of two to guarantee a sufficient length of the FFT.

As shown in FIG. 7, in preparation for channel estimation, a discrete Fourier transform DFT of the cyclic basic midamble code m ^W is performed to obtain a first intermediate result g and Subsequently, the formation of the reciprocal g ⁻¹ of this first intermediate result g takes place. An inverse discrete Fourier transform IDFT is performed on the reciprocal g ⁻¹ and a second intermediate result g ′ is formed. Subsequently, a third intermediate result g ″ is generated by arranging two second intermediate results g ′ together and filling in a vector having the value “zero”. The length of the vector then corresponds to N, the following holds: N ≧ 2L.

Next, the fourth intermediate result g ″ ′ forming the estimation coefficient is represented by the fast Fourier transform FF
Formed from the third intermediate result g "by T. The fourth intermediate result g""is stored in the memory S.
Stored in P.

When the portion of the received signal that depends on the midamble m ^(m) is evaluated, the evaluation is performed according to the following equation: h = IFFT (FFT (em) * g ″ ′).

The vector h then contains the channel coefficients h of, for example, W = 27 of the channel impulse responses of all K connections V1, V2, V3,.

[0072]

(Equation 4)

Here, the solution of the coefficient is 16 bits. From this vector, the channel impulse response with the channel coefficient h can subsequently be extracted for data detection.

The data estimation is effective for the individual data part dt. Furthermore, the interference between the data part dt and the part em of the received signal that depends on the midamble m ^(m) must be taken into account in the data estimation. After the separation of the data symbols d of the data channels DK1 and DK2, demodulation is performed in the demodulator, descrambling is performed in the deinterleaving device, and channel decoding is performed in the convolutional decoder.

On the transmitting side and the receiving side, digital signal processing is controlled by the control device SE. The control device SE is in particular a data channel DK1, D per connection.
Consider the number of K2, the CDMA code of the data channels DK1, DK2, the current radio block structure and the requirements for channel estimation.

In particular, the table T1 is written and read by the control device SE. The table contains the current connections V1, V2, V3 of the radio interface, the midamble code m ^(m) assigned to these connections V1, V2, V3 and the data channels DK1, DK2, DK3 and their CDs.
The MA code is stored.

The mobile radio network combining FDMA, TDMA and CDMA introduced in the embodiment is suitable for the demand for the third generation system. In particular, it is suitable to be implemented in an existing GSM mobile radio network, which requires only a small change cost.

[Brief description of the drawings]

FIG. 1 is a block diagram of a mobile radio network.

FIG. 2 is a schematic diagram of a frame structure of a wireless interface.

FIG. 3 is a schematic diagram of a configuration of a radio block.

FIG. 4 is a block diagram of a transmitter of a wireless station.

FIG. 5 is a block diagram of a receiver of a wireless station.

FIG. 6 is a block diagram of a digital signal processing means.

FIG. 7 is a sequence diagram of channel estimation.

[Procedural Amendment] Submission of translation of Article 34 Amendment of the Patent Cooperation Treaty

[Submission date] June 27, 2000 (2000.6.27)

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] Claims

[Correction method] Change

[Correction contents]

[Claims]

[Procedure amendment 2]

[Document name to be amended] Statement

[Correction target item name] 0004

[Correction method] Change

[Correction contents]

[0004] From DE 195 49 148 a radio communication system using code division multiple access (CDMA) is known, wherein the radio interface additionally has a time division subscriber division. (= Time Division Multiple Access, TDMA = Time Division Multiple Acces
s). In general, in the code multiplexing method, a code sequence (for example, in a form of superposition with a noise signal of a predetermined energy) is added to an individual signal assigned to an individual subscriber, and data to be transmitted by the individual subscriber is transmitted. They are separated. On the other hand, in time division multiplexing (TDMA), time slots that are transmitted in time sequence are allocated to various subscribers, and these time slots are grouped into frames. The time slot sequence is repeated. Further, at the receiving end, a joint detection (JD) method is used to enable improved detection of the transmitted data based on the individual subscriber's knowledge of the CDMA code. In the JD method, CDMA individual signals are commonly captured and detected and supplied to a matched filter. This filter is tuned to the respective individual signal or CDMA code of the individual subscriber, and the output signal of the matched filter is subsequently processed by maximum likelihood decoding to obtain the most likely output signal vector. Can be ascertained. Therefore, the JD method can remove an obstacle of an individual signal due to another individual signal.

[Procedure amendment 3]

[Document name to be amended] Statement

[Correction target item name] 0009

[Correction method] Change

[Correction contents]

Stainer B. et al: “OPTIMUM AND SUBOPTIMUM CHANNEL ESTIMATION FOR THE
UPLINK OD CDMA MOBILE RADIO SYSTEMS WITH JOINT DETECTION ”EUROPEAN TRANS
AKTIONS ON TELECOMMUNICATIONS AND RELATED TECHNOLOGIES, BD 5, NR1, JANUA
R 1994- FEBRUAR 1994, Seiten (page) 39-50, XP000445714 MAILAND, ITALIEN
From N ISSN: 1120-3862, a method for channel estimation in a communication system is known, in which a signal consisting of a data part with data symbols and a midamble with known symbols is transmitted over a transmission channel. On the receiving side, the known symbol signal of the midamble received depending on the channel coefficient and the midamble has a fast Fourier transform or inverse fast Fourier transform having a transform length (number of reference points) according to a power of two. Folded to carry out. SUMMARY OF THE INVENTION It is therefore an object of the invention to provide a method and an apparatus for channel estimation which enable an economical implementation even if the demands on the channel estimation change, in which case the channel estimation is performed at a particularly low cost. I want to be able to.

[Procedure amendment 4]

[Document name to be amended] Statement

[Correction target item name] 0021

[Correction method] Change

[Correction contents]

It is advantageous to assign a plurality of data channels to one connection, wherein a smaller number of midambles is used than the number of data channels. This reduces the cost of channel estimation. Additionally, the number of possible data channels per time slot is increased. Since the data channels use the same midamble and the effect of limiting the capacity of the channel estimation has no effect on the data channel, the number L _M of known symbols of the midamble (m) Is advantageously selected such that L ≧ W * K. Where W is the length of the channel impulse response to be estimated and K is the connection (V1, V2
, V3).

──────────────────────────────────────────────────続 き Continuation of front page (81) Designated country EP (AT, BE, CH, CY, DE, DK, ES, FI, FR, GB, GR, IE, IT, LU, MC, NL, PT, SE ), EA (AM, AZ, BY, KG, KZ, MD, RU, TJ, TM), AU, BR, CA, CN, HU, ID, IL, IN, JP, KR, MX, NO, PL, RU , TR, UA, US, VNF terms (reference) 5K022 DD01 DD33 5K067 AA02 AA03 CC04 CC10 EE02 EE10 JJ11

Claims (12)

[Claims] 1. A channel estimation method in a communication system, comprising: transmitting a signal (e, em) including a data part (dt) having a data symbol (d) and a midamble (m) having a known symbol to a transmission channel; Transmitted through
At the receiving side, the channel coefficient (h) is estimated with respect to the channel impulse response of the transmission channel, and to estimate the channel coefficient (h), the received signal (em) that depends on the midamble (m) is estimated. ) Is cyclically convoluted using estimated coefficients (g ″ ′) derived from known symbols of the midamble (m ^(m) ), and as long as S ≧ 2L holds, an arbitrary number S of reference points Performing a cyclic convolution with: where L is the length of the evaluable part of the midamble (m). 2. The method according to claim 1, wherein the cyclic convolution is performed by a Fourier transform having S reference points. 3. The method of claim 2, wherein the cyclic convolution is performed by a combination of a fast Fourier transform having S reference points and an inverse fast Fourier transform. 4. Estimating the channel coefficient (h) by a plurality of connections (V1, V
2, V3). The method according to any one of claims 1 to 3. 5. The method according to claim 1, further comprising estimating the channel coefficient (h) for transmission channels not used by the connections (V1, V2, V3). 6. The method according to claim 1, wherein the scaling of the estimated channel coefficients is performed. 7. A connection (V1, V2, V3) having one or more data channels (DK1, DK2, DK3) and a connection-specific midamble (m (m)) of the estimated channel coefficient (h). The method according to any one of claims 1 to 6, wherein the correspondence to (1) is performed based on the table (T). 8. The midamble (m ^(m) ) used for channel estimation is matched according to the number K of connections (V1, V2, V3). The described method. 9. The number L _M of known symbols of the midamble (m) is defined as L ≧ W
* K, where W is the length of the channel impulse response to be estimated and K is the number of connections (V1, V2, V3). The described method. 10. The method according to claim 1, wherein the communication system is a mobile radio network and the transmission channels (DK1, DK2, DK3) are radio channels of a radio interface. 11. A channel estimation device in a communication system, comprising: a data part (dt) having a data symbol (d) and a mid part having a known symbol via a transmission channel (DK1, DK2, DK3) of the communication system. A signal (e, em) consisting of an amble (m) and a channel estimator (KS) for estimating a channel coefficient (h) with respect to a channel impulse response of the transmission channel, wherein A channel estimator (KS) converts the received signal (em) that is dependent on the midamble (m) to a midamble (m ^(m) ) for the channel estimator to estimate the channel coefficient (h). Estimated coefficient (g ″ ′) derived from the known symbols of
Is used to convolve cyclically with an arbitrary number S of reference points, where S ≧ 2L holds and L is the length of the evaluable part of the midamble (m). There is a device. 12. The channel estimator (KS) is configured such that the channel estimator performs the estimation of the channel coefficient (h) according to the method of any one of claims 1 to 10. Item 12. The apparatus according to Item 11.