# US 7,778,312 B2 · Method and apparatus for selecting demodulation processing delays in a receiver > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 7,778,312 B2 - Document ID: US7778312B2 - Status: Granted - Kind: B2 - Application number: 11/402,676 - Filed: 2006-04-12 - Published: 2006-08-17 - Granted: 2010-08-17 - Earliest claimed: 2000-12-22 - Assignees: Telefonaktiebolaget LM Ericsson (publ) - Inventors: Douglas A. Cairns; Gregory E. Bottomley - Web page: https://patentgrep.com/patent/US7778312B2 ## CPC - H04L 25/03044: {using fractionally spaced delay lines or combinations of fractionally integrally spaced taps} - H04B 1/712: Weighting of fingers for combining, e.g. amplitude control or phase rotation using an inner loop - H04B 1/7117: Selection, re-selection, allocation or re-allocation of paths to fingers, e.g. timing offset control of allocated fingers - H04B 2201/70707: Efficiency-related aspects - H04B 2201/709727: GRAKE type RAKE receivers - H04L 2025/03477: {not time-recursive} - H04L 2025/03515: {irregularly spaced} - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) - H04B: TRANSMISSION ## Abstract A receiver includes a baseband processor for selecting a set of demodulation processing delays for received signal demodulation from a larger set of candidate delays. In one embodiment, the baseband processor selects the set of demodulation processing delays by calculating at least one metric for each demodulation processing delay in the set of candidate delays, iteratively reducing the set of candidate delays by eliminating one or more demodulation processing delays from the set as a function of comparing the metrics, and setting the processing delays for received signal demodulation to the candidate delays remaining after reduction. In a Generalized RAKE (G-RAKE) embodiment, the metric corresponds to combining weight magnitudes associated with G-RAKE finger delays. In a chip equalizer embodiment, the metric corresponds to coefficient magnitudes associated with equalization filter tap delays. In other embodiments, the metric corresponds to Signal to Interference Ratios (SIRs) associated with the set of candidate delays. ## Claims ### Claim 1 A method of selecting a set of demodulation processing delays for received signal demodulation from a larger set of candidate delays including path delays and other delays, comprising: forming a group of delays including some or all of the candidate delays; calculating a combining weight magnitude for the group of delays based on channel estimates determined for the candidate delays included in the group of delays and impairment correlations between the candidate delays included in the group of delays; iteratively reducing a number of candidate delays included in the group of delays by eliminating one or more of the candidate delays included in the group of delays based on the combining weight magnitude in an initial iteration, in one or more subsequent iterations recalculating the combining weight magnitude to reflect the elimination in a prior iteration of one or more of the candidate delays from the group of delays by removing entries from an impairment covariance matrix that correspond to the one or more candidate delays eliminated in the prior iteration, revising a channel coefficient vector by eliminating vector entries associated with the one or more candidate delays eliminated in the prior iteration and calculating combining weights using the recalculated impairment covariance matrix and the revised channel coefficient vector, and eliminating one or more of the candidate delays remaining in the group of delays based on the combining weight magnitude as recalculated in the current iteration; and setting the processing delays for received signal demodulation to the candidate delays remaining in the group of delays after reduction. ### Claim 2 (depends on 1) The method of claim 1, wherein recalculating an impairment covariance matrix by removing entries from the impairment covariance matrix that correspond to the one or more candidate delays eliminated in the prior iteration comprises subtracting from a portion of an inverse impairment covariance matrix a scaled version of an outer product of a vector associated with each of the candidate delays eliminated in the prior iteration. ### Claim 3 (depends on 2) The method of claim 2, wherein subtracting from the portion of an inverse impairment covariance matrix a scaled version of an outer product of a vector associated with each of the candidate delays eliminated in the prior iteration comprises: modifying the inverse impairment covariance matrix by exchanging rows and columns of the matrix until the last column of the matrix contains entries associated with one of the candidate delays eliminated in the prior iteration; subtracting from a portion of the modified inverse impairment covariance matrix the scaled version of the vector outer product associated with the candidate delay eliminated in the prior iteration; and repeating the operations of modifying and subtracting for each of the candidate delays eliminated in the prior iteration. ### Claim 4 (depends on 3) The method of claim 3, wherein modifying the inverse impairment covariance matrix by exchanging rows and columns of the matrix until the last column of the matrix contains entries associated with one of the candidate delays eliminated in the prior iteration comprises multiplying the inverse impairment covariance matrix with a permutation matrix and a transposed version of the permutation matrix. ### Claim 5 (depends on 1) The method of claim 1, wherein recalculating an impairment covariance matrix by removing entries from the impairment covariance matrix that correspond to the one or more candidate delays eliminated in the prior iteration comprises calculating an N−M square impairment covariance matrix, wherein the N−M square impairment covariance matrix is devoid of any entries associated with M candidate delays eliminated in the prior iteration. ### Claim 6 (depends on 1) The method of claim 1, wherein the combining weights comprise one of Generalized RAKE combining weights or chip equalization filter tap coefficients. ### Claim 7 (depends on 1) The method of claim 1, wherein the demodulation processing delays correspond to one of Generalized RAKE finger delays and equalization filter tap delays. ### Claim 8 (depends on 1) The method of claim 1, comprising, in the initial iteration, eliminating one or more of the candidate delays included in the group of delays having a smallest combining weight calculated in the initial iteration. ### Claim 9 A receiver comprising an antenna and a baseband processor configured to select a set of demodulation processing delays for received signal demodulation from a larger set of candidate delays including path delays and other delays by forming a group of delays including some or all of the candidate delays, calculate a combining weight magnitude for the group of delays based on channel estimates determined for the candidate delays included in the group of delays and impairment correlations between the candidate delays included in the group of delays, iteratively reduce a number of candidate delays included in the group of delays by eliminating one or more of the candidate delays included in the group of delays based on the combining weight magnitude in an initial iteration, in one or more subsequent iterations recalculating the combining weight magnitude to reflect the elimination in a prior iteration of one or more of the candidate delays from the group of delays by removing entries from an impairment covariance matrix that correspond to the one or more candidate delays eliminated in the prior iteration, revising a channel coefficient vector by eliminating vector entries associated with the one or more candidate delays eliminated in the prior iteration and calculating combining weights using the recalculated impairment covariance matrix and the revised channel coefficient vector, and eliminating one or more of the candidate delays remaining in the group of delays based on the combining weight magnitude as recalculated in the current iteration, and set the processing delays for received signal demodulation to the candidate delays remaining in the group of delays after reduction. ### Claim 10 (depends on 9) The receiver of claim 9, wherein the combining weights comprise one of Generalized RAKE combining weights or chip equalization filter tap coefficients. ### Claim 11 (depends on 9) The receiver of claim 9, wherein the demodulation processing delays correspond to one of Generalized RAKE finger delays and equalization filter tap delays. ### Claim 12 (depends on 9) A wireless communication device comprising the receiver according to claim 9. ### Claim 13 (depends on 9) The receiver of claim 9, wherein the baseband processor is configured to iteratively reduce the number of candidate delays included in the group of delays by, in the initial iteration, eliminating one or more of the candidate delays included in the group of delays having a smallest combining weight calculated in the initial iteration.