Method of equalizing a receive data block in a time-division multiple access communication system and receiver utilizing this method

Claim 3

Method according to claim 1 wherein the received block (BD) comprises data symbols (D, D') and a learning sequence (SA), characterized in that said set (P) of parameters comprises parameters (PA) known at the receiving end, including the length of the channel and the length of said learning sequence (SA), and parameters (PC) to be calculated, including the signal to noise ratio and the maximal frequency offset due to the Doppler effect.

Claim 4

Method according to claim 2 characterized in that the parameters (PA) known at the receiving end further comprise the symbol period and the modulus of the symbols.

Claim 5

The method according to claim 2 wherein: the parameters (PC) to be calculated, including the maximal frequency offset due to the Doppler effect, comprise parameters calculated at the receiving end for purposes other than the purpose of equalizing the block (BD).

Claim 6

Method according to claim 1 characterized in that said data block (BD) comprises a learning sequence (SA) including a reference sequence (SR) and precursor symbols and data symbols divided between a first data sequence (D') and a second data sequence (D) respectively transmitted before and after said learning sequence (SA), said communication channel having an impulse response of finite length (K), characterized in that the determination of an optimal length (N opt ) of the data block (BD) includes a stage of optimizing the length (n s , n' s ) of each data sequence (D, D') in which the impulse response of the communication channel is estimated on the basis of the received symbols (R) of the reference sequence (SR) and the precursor symbols, each data sequence (D, D') is associated with an estimate mean-square error, a range for which is determined by respective lower and upper limiting functions (E min (n s ), E max (n s ); E' min (n s ), E' max (n s )) conditioned among other things by the length (n s ) of the data sequence and the length of the reference sequence, and said lower and upper limiting functions are used to determine a range (n opt , n' opt ) for the optimal length of each data sequence corresponding to a specified range of mean-square error.

Claim 7

Method according to claim 5 applied to data blocks comprising constant amplitude zero auto-correlation (CAZAC) or pseudo-CAZAC type sequences, the noise affecting the communication channel being approximated by Gaussian additive white noise, characterized in that the standardized estimate mean-square error corresponding to the data sequence (D) following the reference sequence (SR) is limited by respective lower and upper limit functions (E min (n s ), E max (n s )) as follows: ##EQU33## where: n s is the length of the data sequence, K is the length of the channel, P is the length of the reference sequence, Ts is the symbol period, No/2 is a bilateral Gaussian additive white noise density, E b is the energy transmitted per bit, and f d is the maximum frequency offset due to the Doppler effect, and a range [n 2M , n 1m ] for the optimal length (n opt ) is determined by applying a specified range [ε 1 , ε 2 ] of mean-square error to the lower and upper limiting functions (E min (n s ), E max (n s )).

Claim 8

Method according to claim 5 applied to data blocks comprising constant amplitude zero auto-correlation (CAZAC) and pseudo-CAZAC type sequences, the noise affecting the communication channel being approximated by Gaussian additive white noise, characterized in that the standardized estimate mean-square error corresponding to the data sequence (D) preceding the reference sequence (SR) is limited by lower and upper limit functions (E' min (n' s ), E' max (n' s )) as follows: ##EQU34## where: n' s is the length of the data sequence preceding the reference sequence, K is the length of the channel, P is the length of the reference sequence, T s is the symbol period, No/2 is a bilateral Gaussian additive white noise density, E b is the energy transmitted per bit, and f d is the maximum frequency offset due to the Doppler effect and in that a range for the optimal length (n' opt ) is determined by applying a specified range [ε' 1 , ε' 2 ] of mean-square error to the lower and upper limiting functions (E' min (n' s ), E' max (n' s )).

Claim 9

The method according to claim 1, wherein the equalization process includes the step of accessing the optimal length (Nopt) of the block (BD) by reading one or more truth tables (T) addressed by all or part of the set (P) of parameters.

Claim 10

Method according to claim 8 characterized in that the truth tables (T) supplying the optimal length (Nopt) of the block (BD) are constructed from the results of previous simulations of channel performance.

Claim 11

Method according to claim 8 characterized in that the truth tables (T) for determining the optimal length (Nopt) of the block (BD) are constructed from previous measurements carried out on site.

Claim 12

Method according to claim 8 characterized in that the truth tables (T) for determining the optimal length (Nopt) of the block (BD) are constructed from the results of a theoretical study of receive performance.

Claim 13

The method according to claim 1 applied to constant amplitude zero auto-correlation (CAZAC) or pseudo-CAZAC sequences in which the optimal length (Nopt) of the block (BD) is chosen between a lower limit (n 2M ) and an upper limit (n 1M ), which are determined from receive parameters (P).

Claim 14

A receiver (1) for use in a time-division multiple access (TDMA) communication system, said receiver comprising: means (EG) for equalizing a block of data (BD) transmitted on a communication channel, including means (PO) for tracking variations of said channel; means for estimating, when the block (BD) is received, the impulse response of the channel on which the block (BD) was transmitted; means for evaluating a set (P) of receive parameters; means for providing, according to the set (P) of receive parameters, an optimal length (Nopt) of the block (BD); and means (5) for comparing the optimal length (Nopt) to an actual length (N) of the block (BD); wherein the channel variation tracking means (PO) performs the channel variation tracking to produce an output signal only when the comparator means (5) indicates that the optimal length (Nopt) of the block (BD) is less than the actual length (N) of the block (BD).

Claim 15

A receiver (1) for use in a time-division multiple access (TDMA) communication system, said receiver comprising: means (EG) for equalizing a block of data (BD) transmitted on a communication channel, including means (PO) for tracking variations of said channel; means for estimating, when the block (BD) is received, the impulse response of the channel on which the block (BD) was transmitted; means for evaluating a set (P) of receive parameters; means for providing, according to the set (P) of receive parameters, an optimal length (Nopt) of the block (BD); and means (5) for comparing the optimal length (Nopt) to an actual length (N) of the block (BD); wherein the channel variation tracking means (PO) performs the channel variation tracking to produce an output signal only when the comparator means (5) indicates that the optimal length (Nopt) of the block (BD) is less than the actual length (N) of the block (BD).

Claim 16

Receiver (1) according to claim 13 characterized in that the means for supplying the optimal length (Nopt) of the block (BD) comprise means for storing one or more truth tables (T) addressed by some or all of the receive parameters (P) and supplying as output said optimal length (Nopt) of the block (BD).

Claim 17

Receiver (1) according to claim 13 characterized in that it further comprises means (4) for calculating some receive parameters (PC) including a signal to noise ratio or a maximal frequency offset due to the Doppler effect.

Claim 18

The receiver (1) according to claim 13, wherein the means for providing the optimal length (Nopt) comprises control and processor means.

Claim 19

The receiver according to claim 13 applied to constant amplitude zero auto-correlation (CAZAC) or pseudo-CAZAC sequences in which the optimal length (Nopt) of the block (BD) is chosen between a lower limit (n 2M ) and an upper limit (n 1M ), which are determined from receive parameters (P).