Method and arrangement for detecting the presence of a training signal in a modem receiver

Claim 2

Method of detecting the presence of a periodic training signal in a modem receiver, characterized by the following steps: storing a sequence of the received training signal comprising q·(M+W) samples, where M is the period of the training signal and W is a predetermined window size, both represented as number of symbol intervals T, and q is a sampling factor representing the number of signal samples taken per symbol interval T; determining from two windows comprising two sets of q·W samples and being offset against each other by q·M samples, a metric (z 3 ) in response to the energy contained in the two sample sets and in response to the correlation between the two sample sets, said metric representing the dissimilarity between the two sample sets except for a dissimilarity which is solely due to a carrier frequency offset of the received training signal; and comparing said metric (z 3 ) against a predetermined threshold value, for generating a signal (CYC) indicating the presence of a training signal sequence when the metric falls below said threshold value.

Claim 3

Method according to claim 1, characterized in that for determining the metric (z 3 ), a first auxiliary value (z 1 ) representing the energy of samples contained in the two windows is generated, and a second auxiliary value (z 2 ) representing the correlation between samples in the two windows is generated; and that the quotient of the second auxiliary value and the first auxiliary value is used for generating the metric (z 3 =1-2|z 2 /z 1 ).

Claim 4

Method according to claim 2, characterized in that an estimate for a carrier frequency offset (Δ f ) of the training signal is generated by taking the argument of the second auxiliary value (z 2 ) and dividing it by a given constant (2πMT).

Claim 5

Arrangement for detecting the presence of a received periodic training signal in a modem receiver, characterized in that it comprises: a tapped delay line (21) for q (M+W) samples where M is the number of symbol intervals T representing the length of the training signal period, W is the number of symbol intervals T representing the length of a receiver window, and q is a sampling factor determining the number of samples taken per symbol interval T; first evaluation means (61) responsive to the output signals of two sets of delay line taps (45A, 45B), each set comprising W taps and representing one window (19, 17') of q·W samples, which are offset against each other by q·M taps, for generating a signal (z 1 ) representative of the energy of the signal samples contained in the two windows; second evaluation means (65) responsive to the output signals of the same sets of delay line taps (45A, 45B) as said first evaluation means, for generating a signal (z 2 ) representative of the correlation between the two sets of signal samples contained in the two windows, which signal also constitutes a measure for the carrier frequency offset of the received training signal; third evaluation means (69) responsive to the output signals (z 1 , z 2 ) of said first and second evaluation means for generating a metric (z 3 ) representative of the dissimilarity between the two sets of signal samples in the two windows, except for a difference due to a carrier frequency offset; and fourth evaluation means (73) responsive to the metric (z 3 ) generated by said third evaluation means, for comparing said metric to a predetermined threshold value (8) and for generating on its output a signal indicating the presence of a training signal (CYC) when the value of the metric is below said threshold.

Claim 6

Arrangement for detecting the presence of a received periodic training signal in a modem receiver, characterized in that it comprises: a tapped delay line (21) for q (M+W) samples where M is the number of symbol intervals T representing the length of the training signal period, W is the number of symbol intervals T representing the length of a receiver window, and q is a sampling factor determining the number of samples taken per symbol interval T; first evaluation means (61) responsive to the output signals of two sets of delay line taps (45A, 45B), each set comprising W taps and representing one window (19, 17') of q·W samples, which are offset against each other by q·M taps, for generating a signal (z 1 ) representative of the energy of the signal samples contained in the two windows; second evaluation means (65) responsive to the output signals of the same sets of delay line taps (45A, 45B) as said first evaluation means, for generating a signal (z 2 ) representative of the correlation between the two sets of signal samples contained in the two windows, which signal also constitutes a measure for the carrier frequency offset of the received training signal; third evaluation means (69) responsive to the output signals (z 1 , z 2 ) of said first and second evaluation means for generating a metric (z 3 ) representative of the dissimilarity between the two sets of signal samples in the two windows, except for a difference due to a carrier frequency offset; and fourth evaluation means (73) responsive to the metric (z 3 ) generated by said third evaluation means, for comparing said metric to a predetermined threshold value (8) and for generating on its output a signal indicating the presence of a training signal (CYC) when the value of the metric is below said threshold.

Claim 7

Arrangement according to claim 4, characterized in that said first evaluation means (61) generates as output signal (z 1 ) the sum of the squares of all sample values received on said two sets of delay line taps (45A, 45B).

Claim 8

Arrangement according to claim 4, characterized in that said second evaluation means (65) generates as output signal (z 2 ) the sum of the products of all sample pairs received on corresponding tap pairs (y i , y M+1 * ) of the two sets of delay line taps (45A, 45B).

Claim 9

Arrangement according to claim 5 or 6, characterized in that said third evaluation means (69) determines the value z 3 of its output metric from the values z 1 and z 2 of the signals received from said first (61) and second (65) evaluation means, respectively, by generating z 3 =1-2|z 2 |/z 1 .

Claim 10

Arrangement according to claim 4 or claim 6, characterized in that fifth evaluation means (75) are provided which are connected to the output of said second evaluation means (65), for generating a carrier frequency offset estimation (Δ f ) in response to the output signal (z 2 ) of said second evaluation means by determining the argument (arg z 2 ) of said output signal and dividing it by a given constant (zπMT).

Claim 11

Arrangement according to claim 4, in a modem receiver comprising an equalizer delay line (17) for q·M samples, characterized in that said tapped delay line (21) includes said equalizer delay line (17) as a portion of it.

Claim 12

Arrangement according to claim 4, characterized in that said fourth evaluation means (73') additionally comprises a minimum detection and storing means (77) for detecting and storing a minimum value (z min ) of said metric (z 3 ) generated by said third evaluation means (69) and that it has a first output (43) and a second output (57) providing a copy control signal (COPY) and an averaging control signal (AGV), respectively, said copy control signal being active when the current value of the metric (z 3 ) is below the minimum value (z min ) stored in said minimum detection and storage means (77), and said averaging control signal being active when the current value of the metric (z 3 ) is above the value (z min ) stored in said minimum detection and storage means (77).

Claim 13

Arrangement according to claim 10, characterized in that said fourth evaluation means (73') further comprises counting means (79) for counting the intervals during which the metric (z 3 ) remained below said threshold (Θ), and that when the contents (CYC-CNT) of said counting means (79) has reached a predetermined value (C1), the copying signal (COPY) on said first output (43) is kept inactive and the averaging signal (AGV) on said second output (57) is kept active.

Claim 14

Arrangement according to claim 4 or claim 10 or claim 11, characterized in that it comprises a training sequence buffering and averaging means (47), which has storage for q·M samples and which is connected to q·M output taps (49) of said tapped delay line (21) for receiving a full period (M·T samples) of said train signal, and that in responsive to respective control signals (CYC/COPY, 43; AVG, 57) the training signal buffering and averaging means either stores the sample sequence received on its input lines (49), replacing any previously stored samples; or forms and stores the average of the corresponding samples of a sample sequence stored in it and of any new samples available on its input lines (49).

Claim 15

Arrangement according to claim 12, characterized in that said q·M output taps (49) to which said training sequence buffering and evaluation means (47) is connected are those taps which are symmetrically located with respect to both ends of the tapped delay line (21).