Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04B—TRANSMISSION
  • H04B1/00—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
  • H04B1/69—Spread spectrum techniques
  • H04B1/707—Spread spectrum techniques using direct sequence modulation
  • H04B1/7097—Interference-related aspects
  • H04B1/7103—Interference-related aspects the interference being multiple access interference
  • H04B1/7107—Subtractive interference cancellation
  • H04B1/71075—Parallel interference cancellation
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04B—TRANSMISSION
  • H04B1/00—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
  • H04B1/69—Spread spectrum techniques
  • H04B1/707—Spread spectrum techniques using direct sequence modulation
  • H04B1/709—Correlator structure
  • H04B1/7093—Matched filter type

Definitions

• the present invention relates to digital radiocommunication techniques using multiple access with code distribution (CDMA, “Code-Division Multiple Access”). It relates more particularly to the multi-user detection methods sometimes used in these techniques to improve reception performance.
• CDMA code-Division Multiple Access
• a CDMA signal processed by a receiver has the expression, after filtering and transposition into baseband:
• • sj- ⁇ t) is a generalized code given by the convolution of the impulse response of the u-th channel with the portion corresponding to the symbol bf of the spreading code c u assigned to the channel.
• the number U corresponds to the number of users if each user considered has only one channel. However, there can be more than one channel per user (eg traffic and control).
• the spreading codes c u are sequences of discrete samples called “chips”, with real values ( ⁇ 1) or complex ( ⁇ 1 ⁇ j), having a given chip rate.
• the bf symbols are also with real values ( ⁇ 1) or complex ( ⁇ 1 ⁇ j).
• the duration of a symbol on a channel is a multiple of the duration of the chip, the ratio between the two being the spreading factor Q of the channel.
• the spreading factor may vary from channel to channel.
• a common spreading factor Q equal to the greatest common divisor (PGCD) of the U spreading factors Q u .
• PGCD common divisor
• the duration of the generalized response s u (t) corresponds to Q + W-1 chips if
• W denotes the length of the impulse response expressed in number of chips.
• N is a random noise vector of size n ⁇ Q + W-1;
• A (A 1 , A 2 A n ) is a matrix of generalized size codes
• (1 ⁇ u ⁇ U) is a convolution of the impulse response of the u-th channel and of the Q samples of the spreading code of the u-th channel corresponding to the i-th symbol of the block.
• Mf is a Toeplitz matrix of size (n ⁇ Q + W-1) x (n ⁇ Q + WQ) obtained from the values c ⁇ q) of the chips of the spreading code c u of the u-th channel for the duration of i -th bit of the block: and f is a column vector of size (n-1) ⁇ Q + W which, when the U channels are received in synchronized manner, contains (i-1) ⁇ Q components at zero, followed by the W samples of the impulse response of the u-th channel relative to the i-th bf symbol, followed by (ni) ⁇ Q other components at zero.
• the most commonly used receiver uses one or more suitable filters to estimate the value of the symbols transmitted on each channel. This receiver estimates the impulse response of the channel according to a propagation path or several propagation paths ("rake receiver").
• the nxU components Zf of the vector Z are respective flexible estimates of the nxU symbols bj- 1 of the vector b. If the decodings carried out downstream admit flexible estimates as input, the components of the vector Z can be used directly. Otherwise, the sign of these components is taken to form the hard estimates of the symbols.
• the adapted filter receiver is optimal when the generalized codes (vectors ⁇ f) are orthogonal two by two, that is to say when the matrix
• a * .A is diagonal.
• systems adopt spreading codes orthogonal two by two and having good autocorrelation properties, which makes it possible to verify this condition as a first approximation.
• An object of the present invention is to compensate for the non-optimal nature of the filter receiver adapted in these situations.
• the invention thus proposes a method for receiving a radio signal comprising contributions from several multiplex channels by respective spreading codes, in which a filter receiver adapted to each multiplex channel is allocated to estimate an impulse response of the channel and provide flexible first estimates of symbols transmitted on the channel.
• the first flexible estimates provided by the receivers with suitable filters are processed to obtain a corrected flexible estimate of at least one symbol transmitted on a channel by subtracting from the first flexible estimate of said symbol at least one term equal to the product.
• the method corrects the soft estimates provided by the matched filter receiver taking into account the particular form of interference caused in a channel by the presence of the other multiple access channels. This interference is here called MAI (“Multiple Access Interference”). This taking into account improves the performance of the receiver in terms of signal-to-noise ratio.
• MAI Multiple Access Interference
• the symbols are typically transmitted over U multiplex channels in the form of respective blocks of n symbols, n and U being numbers greater than 1.
• Q is the number of samples per symbol in the spreading codes
• W is the number of samples in the impulse response estimates
• R j 0 -b j represents, as a first approximation, an estimate of the interference caused in the u-th channel by the presence of the U-1 other channels. This approximation may suffice in cases where inter-symbol interference (ISI) is low.
• ISI inter-symbol interference
• ISI inter-symbol interference
• the corrected flexible estimates will be determined sequentially for several symbols of the U blocks.
• the flexible estimates corrected for said symbols are then advantageously taken if they have been previously determined, and the first flexible estimates for said symbols otherwise.
• Another aspect of the present invention relates to a device for receiving a radio signal comprising contributions from several multiplex channels by respective spreading codes, comprising receptors with suitable filters each assigned to a respective multiplex channel to estimate a impulse response of the channel and providing first flexible estimates of symbols transmitted on the channel, and means for processing these first flexible estimates to obtain a corrected flexible estimate of at least one symbol transmitted on a channel as previously indicated.
• FIG. 1 is a block diagram of a reception device according to the invention
• - Figure 2 is a diagram of a filter receiver adapted from the device.
• the device shown in FIG. 1 is part of the reception stage of a radiocommunication station capable of communicating with several remote stations 1.
• the uplink channels used by these remote stations 1 are multiplexed by the CDMA technique, so that the radio signal received by antenna 2, reduced to baseband, can be represented in the form (1) - (2) for U multiplex channels from V stations (1 ⁇ V ⁇ U).
• the station incorporating the device is for example a base station of a third generation cellular radiocommunication system of UMTS type ("Universal Mobile Telecommunication System").
• the unit 3 schematically represents the modules conventionally carrying out the signal reception pre-processing (amplification, filtering, conversion to base band, sampling at the frequency of the chips).
• This unit 3 delivers blocks Y of n ⁇ Q + W-1 samples, corresponding to blocks of n symbols transmitted simultaneously on the U channels. If the blocks of n symbols succeed one another without interruption on the channels, there is an overlap of W samples (chips) between the successive blocks Y, corresponding to the duration of the impulse response.
• the received signal blocks Y are supplied in parallel to U receivers with suitable filters 4 U operating with respective channel codes c u produced by pseudo-random code generators 5 U (1 ⁇ u ⁇ U).
• FIG. 2 illustrates the well-known structure of a 4 U matched filter receiver of the “rake” type.
• This receiver 4 U includes a channel probing unit 6 which evaluates the impulse response of the u-th channel by searching for K propagation paths (K> 1), for the K “fingers” of the receiver.
• K propagation paths
• Each path k is characterized by a delay t expressed in number of chips and a complex response r u (1 ⁇ k ⁇ K).
• the signal transmitted on each channel by a remote station may include sequences of known training symbols.
• the unit 6 By searching, over a window of length W chips, for the K correlations of greatest amplitude between the received signal Y and these known sequences modulated by the spreading code c u of the channel, the unit 6 obtains the delays t ⁇ ⁇ ( time shifts of the maxima) and the responses r ⁇ (maximum values).
• the spreading code c u produced by the generator 5 U (or its conjugate if the codes are complex) is delayed by a unit 7 which applies to it the delay of t ⁇ ⁇ chips.
• Each code thus delayed is multiplied by the received signal Y (multiplier 8) and by the conjugate of the complex response r u (multiplier 9).
• the K results of these multiplications are added by a summator 10 to form the block Z u of n flexible estimates for the u-th channel.
• the i-th component of the block Z u is the flexible estimation of the symbol b ⁇ . If the symbols bf are signed bits ( ⁇ 1), the flexible estimates of the block Z u are the real parts of the summed contributions of the K fingers. If the symbols b ⁇ are pairs of signed bits, they are complex numbers equal to these summed contributions.
• the flexible estimates Z u can be transformed into hard estimates b ⁇ by decision modules 12 u at the output of the receivers 4 U.
• the modules 12 u simply apply the sign function to the real components of the vectors Z u .
• the modules 12 u apply the sign function to the real parts and to the imaginary parts of the components of the vectors Z u .
• Each matrix R,. • of size U x U contains the correlations of the generalized codes between the i-th symbols and the (i + j) -th symbols of the blocks relating to the U channels.
• equation (8) is reduced to:
• Z j is a vector of size U containing the flexible estimates of the i-th symbols of the U blocks and N ( a corresponding noise vector.
• the method according to the invention comprises a postprocessing of the flexible estimates of the vectors Z j , which is carried out in the module 13 represented in FIG. 1.
• the algorithm used is called MFPIC (“Matched Filter Parallel Interference Cancellation”).
• Relations (10) show that the quantity of computations required can be reduced thanks to the properties of symmetry of the matrices R,,.
• Each vector ⁇ j U contains the convolution of the estimated impulse response of channel u and the Q samples of the spreading code of this channel corresponding to the i-th symbol of the block, and is defined as in relation (5), the matrix M- ⁇ being determined according to (6) according to the code c u supplied by the generator 5 U , and the vector H ⁇ being replaced by an estimated response vector H j U containing the complex responses r ⁇ estimated by the units of hole 6, positioned according to the corresponding delays tj ⁇ .
• the matrix R j 0 is equal to the correlation matrix R j 0 in which the diagonal components are set to zero.
• the block of n symbols is preceded and followed by other symbols, the hard estimates of which are placed in the vectors b,; for ij ⁇ 1 and b i + : for i + j> n. Otherwise, these vectors can be set to zero.
• the vectors X j thus obtained are flexible estimates corrected taking into account relations (11). This correction uses the decisions made in hard estimates b j , and therefore a certain structure of MAI and NSI, which is not the same as that of Gaussian noise N.
• the flexible estimates of the vectors X j and / or the hard estimates of the vectors b j (1 ⁇ i ⁇ n) are
• the module 13 sequentially executes the following operations (13) and (14) for i ranging from 1 to n:
• the corrected estimates X j which have already been calculated are taken into account recursively in the decision taken in operation (14), which further improves the estimates.
• the operations (13) and (14) above could be executed in an order other than that of the increasing indexes i. For example, we could execute them in an order determined according to an energy criterion.
• One possibility is to correct first the estimates of the least energetic symbols in the received signal, that is to say to proceed in the order of the indexes i for which the diagonal terms of the correlation matrix R- 0 are decreasing.
• the function applied (in decision modules 12 u ) to deduce flexible estimates Z the estimates bj used in formula (13) is, rather than the sign function, a generally increasing function between -1 and +1.
• a three-valued function (-1 for Zf ⁇ -T, 0 for -T ⁇ Z, ⁇ ⁇ + T and +1 for Zf> + T) eliminates the need to make corrections based on low likelihood estimates relative to a threshold T.
• the function can also grow continuously from -1 to +1.
• the advantage of the sign function is mainly in terms of complexity since it avoids multiplication in the formula
• the MFPIC algorithm offers multi-user detection with good performance, particularly for the relatively low spreading factors Q. As soon as the bit error rate is less than 15%, it provides a significant gain in terms of signal-to-noise ratio, compared to the simple rake receiver. Its limitations seem to come only from uncertainties in the estimates of the impulse responses of the channels.

Landscapes

• Engineering & Computer Science (AREA)
• Computer Networks & Wireless Communication (AREA)
• Signal Processing (AREA)
• Noise Elimination (AREA)
• Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)
• Mobile Radio Communication Systems (AREA)

Applications Claiming Priority (3)

Publications (1)

Family

ID=8851562

Family Applications (1)

Country Status (5)

Families Citing this family (21)

Family Cites Families (8)

• 2000
  • 2000-06-22 FR FR0008008A patent/FR2810820B1/fr not_active Expired - Fee Related
• 2001
  • 2001-06-20 WO PCT/FR2001/001934 patent/WO2001099301A1/fr not_active Ceased
  • 2001-06-20 US US10/312,058 patent/US7167529B2/en not_active Expired - Lifetime
  • 2001-06-20 EP EP01947548A patent/EP1293048A1/de not_active Withdrawn
  • 2001-06-20 AU AU2001269209A patent/AU2001269209A1/en not_active Abandoned

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events