# US 8,787,427 B2 · Chip-level processing for joint demodulation in CDMA receivers > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 8,787,427 B2 - Document ID: US8787427B2 - Status: Granted - Kind: B2 - Application number: 13/468,680 - Filed: 2012-05-10 - Published: 2013-11-14 - Granted: 2014-07-22 - Earliest claimed: 2012-05-10 - Assignees: Telefonaktiebolaget L M Ericsson (publ) - Applicants: Andres Reial; Stephen Grant; Michael Samuel Bebawy; Yi-Pin Eric Wang - Inventors: Andres Reial; Stephen Grant; Michael Samuel Bebawy; Yi-Pin Eric Wang - Web page: https://patentgrep.com/patent/US8787427B2 ## CPC - H04B 1/7105: Joint detection techniques, e.g. linear detectors - H04B 1/7097: Interference-related aspects - H04B 1/71052: {using decorrelation matrix} - H04B 1/71057: {using maximum-likelihood sequence estimation [MLSE]} - H04L 25/03331: {Arrangements for the joint estimation of multiple sequences} - H04B: TRANSMISSION - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) ## Abstract A CDMA multi-code joint demodulation solution in which impairment suppression and channel matching operations are performed prior to despreading. Embodiments include a linear front end that performs chip-level suppression of signal components that are not included in a subsequent joint demodulation process. The pre-processing stage also carries out metric preparation and provides a vector decision statistic that is processed by a joint demodulation stage to extract per-code soft values for the symbols of interest in the received signal. Both code-specific and code-averaged versions of the linear processing are disclosed, as are several front-end configurations with equivalent performance, but different complexity trade-offs. These new approaches use a block formulation, requiring a set of input chip samples as an input, and perform all operations as matrix-vector multiplications, which is an approach amenable to efficient DSP or hardware implementation. ## Claims ### Claim 1 A method for processing a received spread spectrum signal containing two or more simultaneously transmitted symbols of interest with known spreading codes and one or more interfering signals with known spreading codes, wherein the known spreading codes for the simultaneously transmitted symbols of interest differ from one another and from the known spreading code or codes for the interfering signals, said method comprising: for a given despread symbol interval, forming a sample vector from samples of the received spread spectrum signal, wherein said samples span the despread symbol interval; using one or more matrix multiplication operations, performing a linear transformation of the sample vector to an interference-reduced decision statistic vector for the given despread symbol interval, based on covariance statistics for the received signal, wherein said linear transformation depends on the known spreading codes and wherein the interference-reduced decision statistic vector comprises a single complex value for at least each of the symbols of interest; and estimating a symbol value for at least one of the symbols of interest from the interference-reduced decision statistic vector, for the given despread symbol interval, using a non-linear detection process that accounts for code-specific leakage between the known spreading codes for the symbols of interest. ### Claim 2 (depends on 1) The method of claim 1, further comprising computing an impairment covariance matrix representing an estimation of impairment cross-correlations among the sample vector samples, wherein said linear transformation of the sample vector is based on the computed impairment covariance matrix. ### Claim 3 (depends on 2) The method of claim 2, wherein computing the impairment covariance matrix comprises: calculating a first term corresponding to code-specific, same-symbol interference from one or more of the interfering signals having known spreading codes; calculating one or more additional terms corresponding to code-specific inter-symbol interference from at least the known spreading codes for the symbols of interest; and summing at least the first term and the one or more additional terms to form the impairment covariance matrix. ### Claim 4 (depends on 2) The method of claim 2, wherein computing the impairment covariance matrix comprises calculating a first term corresponding to code-averaged, same-symbol interference from one or more of the interfering signals having known spreading codes. ### Claim 5 (depends on 4) The method of claim 4, wherein computing the impairment covariance matrix further comprises: calculating one or more additional terms corresponding to code-averaged inter-symbol interference from at least known spreading codes for the symbols of interest; and summing at least the first term and the one or more additional terms to form the impairment covariance matrix. ### Claim 6 (depends on 1) The method of claim 1, further comprising computing a signal covariance matrix representing an estimation of correlations among the sample vector samples, wherein said linear transformation of the sample vector is based on the computed signal covariance matrix. ### Claim 7 (depends on 1) The method of claim 1, wherein estimating the symbol value for at least one of the symbols of interest from the interference-reduced decision statistic vector comprises: computing a joint detection metric from each of a plurality of symbol vector hypotheses, where each symbol vector hypothesis comprises a candidate symbol for each of the symbols of interest; and performing a tree search, using the joint detection metrics, to jointly detect symbol values for the symbols of interest. ### Claim 8 (depends on 7) The method of claim 7, wherein said tree search is a breadth-first constrained tree search, or a depth-first constrained tree search, or a combination of both. ### Claim 9 (depends on 7) The method of claim 7, wherein said tree search is a reduced complexity tree search approach selected from a group consisting of: an m-algorithm search; a sphere-decoding search; a geometric-decoding search; and an assisted maximum likelihood detection (AMLD) search. ### Claim 10 (depends on 1) The method of claim 1, wherein said non-linear detection process comprises estimating a symbol value for each symbol among the symbols of interest from the interference-reduced decision statistic vector, for the given symbol interval, based on one or more decision metrics, each of which is determined by a joint symbol-value hypothesis and code-specific leakage associated with one or more symbols of interest. ### Claim 11 A receiver circuit adapted to process a received spread spectrum signal containing two or more simultaneously transmitted symbols of interest with known spreading codes and one or more interfering signals with known spreading codes, wherein the known spreading codes for the simultaneously transmitted symbols of interest differ from one another and from the known spreading code or codes for the interfering signals, the receiver circuit comprising: a front-end radio frequency circuit configured to mix the received spread spectrum signal to baseband and to digitize the received spread spectrum signal; and a receiver processing circuit configured to: form a sample vector for a given despread symbol interval, from samples of the received spread spectrum signal, wherein said samples span the despread symbol interval; perform a linear transformation of the sample vector to an interference-reduced decision statistic vector for the given despread symbol interval, using one or more matrix multiplication operations, based on covariance statistics for the received signal, wherein said linear transformation depends on the known spreading codes and wherein the interference-reduced decision statistic vector comprises a single complex value for at least each of the symbols of interest; and estimate a symbol value for at least one of the symbols of interest from the interference-reduced decision statistic vector, for the given despread symbol interval, using a non-linear detection process that accounts for code-specific leakage between the known spreading codes for the symbols of interest. ### Claim 12 (depends on 11) The receiver circuit of claim 11, wherein said receiver processing circuit is further configured to compute an impairment covariance matrix representing an estimation of impairment cross-correlations among the sample vector samples, wherein said linear transformation of the sample vector is based on the computed impairment covariance matrix. ### Claim 13 (depends on 12) The method of claim 12, wherein the receiver processing circuit is configured to compute the impairment covariance matrix by calculating a first term corresponding to code-specific, same-symbol interference from one or more of the interfering signals having known spreading codes. ### Claim 14 (depends on 13) The method of claim 13, wherein the receiver processing circuit is further configured to compute the impairment covariance matrix by: calculating one or more additional terms corresponding to code-specific inter-symbol interference from at least the known spreading codes for the symbols of interest; and summing at least the first term and the one or more additional terms to form the impairment covariance matrix. ### Claim 15 (depends on 12) The method of claim 12, wherein the receiver processing circuit is configured to compute the impairment covariance matrix by: calculating a first term corresponding to code-averaged, same-symbol interference from one or more of the interfering signals having known spreading codes; calculating one or more additional terms corresponding to code-averaged inter-symbol interference from at least known spreading codes for the symbols of interest; and summing at least the first term and the one or more additional terms to form the impairment covariance matrix. ### Claim 16 (depends on 11) The receiver circuit of claim 11, wherein said receiver processing circuit is further configured to compute a signal covariance matrix representing an estimation of correlations among the sample vector samples, wherein said linear transformation of the sample vector is based on the computed signal covariance matrix. ### Claim 17 (depends on 11) The receiver circuit of claim 11, wherein the receiver processing circuit is configured to estimate the symbol value for at least one of the symbols of interest from the interference-reduced decision statistic vector by: computing a joint detection metric from each of a plurality of symbol vector hypotheses, where each symbol vector hypothesis comprises a candidate symbol for each of the symbols of interest; and performing a tree search, using the joint detection metrics, to jointly detect symbol values for the symbols of interest. ### Claim 18 (depends on 17) The receiver circuit of claim 17, wherein said tree search is a breadth-first constrained tree search, or a depth-first constrained tree search, or a combination of both. ### Claim 19 (depends on 17) The receiver circuit of claim 17, wherein said tree search is a reduced complexity tree search approach selected from a group consisting of: an m-algorithm search; a sphere-decoding search; a geometric-decoding search; and an assisted maximum likelihood detection (AMLD) search. ### Claim 20 (depends on 11) The receiver circuit of claim 11, wherein said non-linear detection process comprises the estimation of a symbol value for each symbol among the symbols of interest from the interference-reduced decision statistic vector, for the given symbol interval, based on one or more decision metrics, each of which is determined by a joint symbol-value hypothesis and code-specific leakage associated with one or more symbols of interest.