Method and apparatus for channel characterization in direct sequence spread spectrum based wireless communication systems

A system for characterization of a multi-path channel in a Direct Sequence Spread Spectrum (DSSS) based wireless communication system, the multi-path channel being modeled as a tapped delay line Finite Impulse Response (FIR) filter, the tapped delay line FIR filter having L taps, each of the L taps of the tapped delay line FIR filter representing a path of the multi-path channel, each of the L taps having a tap coefficient associated with it, the channel being modeled by estimating the coefficients of each of the L taps, each of the L tap coefficients being estimated by solving a set of L simultaneous linear equations, the set of L simultaneous linear equations being obtained by using a received signal, the signal received at a receiver is a transmitted spread pre-defined training sequence on the multi-path channel, a spread pre-defined training sequence is obtained by spreading each symbol of a pre-defined training sequence by a pre-defined spreading sequence, the pre-defined training sequence being any known sequence of symbols, the samples of the received signal being used to estimate a symbol boundary, the symbol boundary being a time of arrival of a symbol in a strongest path of the multi-path channel, the estimated symbol boundary being used to estimate a time of arrival of an earliest path of the multi-path channel, regardless of the energy in the earliest path, the system comprising:

• a buffer buffering 3N samples of the received signal, the buffered samples covering three consecutive symbols of the received signal, a symbol lying in middle of the three consecutive symbols being buffered at a 0th time lag, the 0th time lag corresponding to the estimated symbol boundary;
• b. a correlator computing 2L−1 cross correlations between the buffered 3N samples and the pre-defined spreading sequence, for a pre-defined set of 2L−1 time lags, the pre-defined set of 2L−1 time lags covering L−1 positive time lags and L−1 negative time lags with respect to the 0th time lag, the 2L−1 correlations are estimated symmetrically around the 0th time lag to capture pre-cursor and post-cursor parts of the multi-path channel impulse response;
• c. a set selector selecting L possible sets of L cross correlations for L consecutive time lags, from the computed 2L−1 cross correlations;
• d. an energy calculator computing the energy in each of the L possible sets, the energy being computed by adding the squares of the magnitudes of each of the L cross correlations in each of the L possible sets;
• e. a peak selector selecting a set of L cross correlations having the maximum energy among all the L possible sets of L cross correlations, the selected set of L cross correlations covering maximum pre-cursor and post-cursor energy, a time instant corresponding to the time lag of the first element in the selected set of L cross correlations corresponding to the time of arrival of the earliest path of the multi-path channel, regardless of the energy in the earliest path;
• and
• f. a tap coefficient estimator estimating each of the L tap coefficients by using the selected set of L cross correlations, forming a set of L simultaneous linear equations by using the selected set of L cross correlations, the three consecutive symbols of the received signal, and autocorrelations of the pre-defined spreading sequence, wherein the non-diagonal elements of the autocorrelation matrix need not be zero, thereby, the pre-defined spreading sequence need not be orthogonal with each other at different time lags, determining each of the L tap coefficients by solving the set of L simultaneous linear equations, the each of the L tap coefficients being represented in a matrix form h, wherein h=Rbb−1Rby, and h denotes a vector [h0 h1 h2 . . . h(L−1)]T, with hi denoting ith tap coefficient, Rby denoting the vector [p0 p1 p2 . . . p(L−1)]T, with pi denoting ith cross correlation in the selected set of L cross correlations, and Rbb denoting an autocorrelation matrix.

A computer program product disposed on a computer readable storage medium storing instructions that, when executed by a computer, cause the computer to perform characterization of a multi-path channel in a Direct Sequence Spread Spectrum (DSSS) based wireless communication system, the multi-path channel being modeled as a tapped delay line Finite Impulse Response (FIR) filter, the tapped delay line FIR filter having L taps, each of the L taps of the tapped delay line FIR filter representing a path of the multi-path channel, each of the L taps having a tap coefficient associated with it, the channel being modeled by estimating the coefficients of each of the L taps, each of the L tap coefficients being estimated by solving a set of L simultaneous linear equations, the set of L simultaneous linear equations being obtained by using a received signal, the signal received at a receiver is a transmitted spread pre-defined training sequence on the multi-path channel, a spread pre-defined training sequence is obtained by spreading each symbol of a pre-defined training sequence by a pre-defined spreading sequence, the pre-defined training sequence being any known sequence of symbols, the samples of the received signal being used to estimate a symbol boundary, the symbol boundary being a time of arrival of a symbol in a strongest path of the multi-path channel, the estimated symbol boundary being used to estimate a time of arrival of an earliest path regardless of the energy in the earliest path of the multi-path channel, the computer program product comprising instructions to perform the steps of:

• a. buffering 3N samples of the received signal, the buffered samples covering three consecutive symbols of the received signal, a symbol lying in middle of the three consecutive symbols being buffered at a 0th time lag, the 0th time lag corresponding to the estimated symbol boundary;
• b. computing 2L−1 cross correlations between the buffered 3N samples and the pre-defined spreading sequence, for a pre-defined set of 2L−1 time lags, the pre-defined set of 2L−1 time lags covering L−1 positive time lags and L−1 negative time lags with respect to the 0th time lag, the 2L−1 correlations are estimated symmetrically around the 0th time lag to capture pre-cursor and post-cursor parts of the multi-path channel impulse response;
• c. selecting L possible sets of L cross correlations for L consecutive time lags, from the computed 2L−1 cross correlations;
• d. computing the energy in each of the L possible sets, the energy being computed by adding the squares of the magnitudes of each of the L cross correlations in each of the L possible sets;
• e. selecting a set of L cross correlations having the maximum energy among all the L possible sets of L cross correlations, the selected set of L cross correlations covering maximum pre-cursor and post-cursor energy, a time instant corresponding to the time lag of the first element in the selected set of L cross correlations corresponding to the time of arrival of the earliest path of the multi-path channel, regardless of the energy in the earliest path;
• and
• f. estimating each of the L tap coefficients by using the selected set of L cross correlations, wherein the step of estimating each of the L tap coefficients includes the steps of:
• i. forming a set of L simultaneous linear equations by using the selected set of L cross correlations, the three consecutive symbols of the received signal, and autocorrelations of the pre-defined spreading sequence, wherein the non-diagonal elements of the autocorrelation matrix need not be zero, thereby, the pre-defined spreading sequence need not be orthogonal with each other at different time lags;
• and
• ii. determining each of the L tap coefficients by solving the set of L simultaneous linear equations, the each of the L tap coefficients being represented in a matrix form h, wherein h=Rbb−1Rby, and h denotes a vector [h0 h1 h2 . . . h(L−1)]T, with hi denoting the ith tap coefficient Rby denoting the vector [p0 p1 p2 . . . p(L−1)]T, with pi denoting ith cross correlation in the selected set of L cross correlations, and Rbb denoting an autocorrelation matrix.