# US 2014/0153678 A1 · Estimating Optimal Linear Regression Filter Length for Channel Estimation > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2014/0153678 A1 - Document ID: US20140153678A1 - Status: Published application - Kind: A1 - Application number: 13/693,561 - Filed: 2012-12-04 - Published: 2014-06-05 - Earliest claimed: 2012-12-04 - Assignees: Telefonaktiebolaget LM Ericsson (publ) - Inventors: Elias Jonsson; Andres Reial - Web page: https://patentgrep.com/patent/US20140153678A1 ## CPC - H04L 25/0222: {Estimation of channel variability, e.g. coherence bandwidth, coherence time, fading frequency} - H04L 25/0202: {Channel estimation} - H04L 25/0204: {of multiple channels} - H04L 25/022: {of frequency response} - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) ## Abstract Techniques for obtaining smoothed channel response estimates in a communications receiver. An example method begins with obtaining ordered channel measurement samples for a current estimation interval and continues with the estimation of a short-term slope value for each of several sequential observation time intervals. A distortion metric is calculated for each of a plurality of candidate filter lengths, where each candidate filter length is an integer multiple of the uniform length. The distortion metric is based on the short-term slope values for observation time intervals falling within the candidate filter length, and characterizes the extent to which the candidate filter length will introduce tracking bias in subsequent signal processing. A filter length is then selected for use in demodulating a received signal from at least among the candidate filter lengths, based on the distortion metric values. ## Claims ### Claim 1 A method, in a communications receiver, for obtaining smoothed channel response estimates, the method comprising: performing multiple channel response measurements corresponding to a path delay or to a signal frequency, to obtain ordered channel measurement samples for a current estimation interval; estimating a short-term slope value for each of a plurality of sequential observation time intervals, each observation time interval comprising a plurality of the ordered channel measurement samples and having a uniform length; for each of a plurality of candidate filter lengths, where each candidate filter length is an integer multiple of the uniform length, computing a distortion metric value based on the short-term slope values for observation time intervals falling within the candidate filter length, the distortion metric value characterizing an extent to which the candidate filter length will introduce a tracking bias in filtering the ordered channel measurement samples; and selecting a filter length for use in demodulating a received signal from at least among the candidate filter lengths, based on the distortion metric values, such that the longest filter length meeting a pre-determined distortion criterion is selected. ### Claim 2 (depends on 1) The method of claim 1, wherein selecting the filter length from at least among the candidate filter lengths comprises selecting the filter length from among the candidate filter lengths and a default filter length, said default filter length being shorter than any of the candidate filter lengths. ### Claim 3 (depends on 1) The method of claim 1, wherein estimating a short-term slope value comprises determining a best-fit line for the ordered channel measurement samples of the observation time interval, using a linear regression model. ### Claim 4 (depends on 1) The method of claim 1, wherein computing the distortion metric value for each candidate filter length comprises determining minimum and maximum slope values among the short-term slope values for observation time intervals falling within the candidate filter length and normalizing a difference between the minimum and maximum slope values. ### Claim 5 (depends on 4) The method of claim 4, wherein the distortion metric value p for each candidate filter length is computed according to the formula p = α re , max - α re , min + α im , max - α im , min max  (  α re , max  ,  α re , min  ) + max  (  α im , max  ,  α im , min  ) , where αre,max and αre,min are the maximum and minimum values of the real part of the short-term slope values, respectively, and αim,max and αim,min are the maximum and minimum values of the imaginary part of the short-term slope values, respectively. ### Claim 6 (depends on 1) The method of claim 1, wherein computing the distortion metric value for each candidate filter length comprises calculating a variance over the short-term slope values for observation time intervals falling within the candidate filter length. ### Claim 7 (depends on 1) The method of claim 1, further comprising filtering multiple distortion metric values for each candidate filter length to obtain filtered distortion metric values, wherein said selecting the filter length is based on the filtered distortion metric values. ### Claim 8 (depends on 7) The method of claim 7, wherein selecting the filter length from at least among the candidate filter lengths comprises comparing the filtered distortion metric for the selected candidate filter length to a scaled distortion metric for a pre-determined reference filter length. ### Claim 9 (depends on 8) The method of claim 8, wherein said pre-determined distortion criterion is that the filtered distortion metric is less than or equal to the scaled distortion metric. ### Claim 10 (depends on 8) The method of claim 8, further comprising adapting a scaling factor for the scaled distortion metric based on one or more signal quality measurements. ### Claim 11 (depends on 1) The method of claim 1, wherein the ordered channel measurement samples are time-ordered channel measurement samples taken over an interval of time. ### Claim 12 (depends on 1) The method of claim 1, wherein the ordered channel measurement samples are frequency-ordered channel measurement samples taken across a set of sub-carrier frequencies. ### Claim 13 (depends on 1) The method of claim 1, further comprising repeating each of the performing, estimating, computing, and selecting operations for each of a series of channel estimation intervals. ### Claim 14 A receiver circuit comprising a channel response measurement circuit configured to perform a plurality of channel response measurements corresponding to a path delay or to a signal frequency, to obtain ordered channel measurement samples for a current estimation interval, and a processing circuit configured to: estimate a short-term slope value for each of a plurality of sequential observation time intervals, each observation time interval comprising a plurality of the ordered channel measurement samples and having a uniform length; for each of a plurality of candidate filter lengths, where each candidate filter length is an integer multiple of the uniform length, compute a distortion metric value based on the short-term slope values for observation time intervals falling within the candidate filter length, the distortion metric value characterizing an extent to which the candidate filter length will introduce a tracking bias in filtering the ordered channel measurement samples; and select a filter length for use in demodulating a received signal from at least among the candidate filter lengths, based on the distortion metric values, such that the longest filter length meeting a pre-determined distortion criterion is selected. ### Claim 15 (depends on 14) The receiver circuit of claim 14, wherein the processing circuit is configured to select the filter length from at least among the candidate filter lengths by selecting the filter length from among the candidate filter lengths and a default filter length, said default filter length being shorter than any of the candidate filter lengths. ### Claim 16 (depends on 14) The receiver circuit of claim 14, wherein the processing circuit is configured to estimate short-term slope values by determining best-fit lines for the ordered channel measurement samples of the observation time interval, using a linear regression model. ### Claim 17 (depends on 14) The receiver circuit of claim 14, wherein the processing circuit is configured to compute the distortion metric value for each candidate filter length by determining minimum and maximum slope values among the short-term slope values for observation time intervals falling within the candidate filter length and normalizing a difference between the minimum and maximum slope values. ### Claim 18 (depends on 17) The receiver circuit of claim 17, wherein the processing circuit is configured to compute the distortion metric value p for each candidate filter length according to the formula p = α re , max - α re , min + α im , max - α im , min max  (  α re , max  ,  α re , min  ) + max  (  α im , max  ,  α im , min  ) , where αre,max and αre,min are the maximum and minimum values of the real part of the short-term slope values, respectively, and αim,max and αim,min are the maximum and minimum values of the imaginary part of the short-term slope values, respectively. ### Claim 19 (depends on 14) The receiver circuit of claim 14, wherein the processing circuit is configured to compute the distortion metric value for each candidate filter length by calculating a variance over the short-term slope values for observation time intervals falling within the candidate filter length. ### Claim 20 (depends on 14) The receiver circuit of claim 14, wherein the processing circuit is further configured to filter multiple distortion metric values for each candidate filter length to obtain filtered distortion metric values, and to select the filter length based on the filtered distortion metric values. ### Claim 21 (depends on 20) The receiver circuit of claim 20, wherein the processing circuit is configured to select the filter length from at least among the candidate filter lengths by comparing the filtered distortion metric for the selected candidate filter length to a scaled distortion metric for a pre-determined reference filter length. ### Claim 22 (depends on 21) The receiver circuit of claim 21, wherein said pre-determined distortion criterion is that the filtered distortion metric is less than or equal to the scaled distortion metric. ### Claim 23 (depends on 22) The receiver circuit of claim 22, wherein the processing circuit is further configured to adapt a scaling factor for the scaled distortion metric based on one or more signal quality measurements. ### Claim 24 (depends on 14) The receiver circuit of claim 14, wherein the ordered channel measurement samples are time-ordered channel measurement samples taken over an interval of time. ### Claim 25 (depends on 14) The receiver circuit of claim 14, wherein the ordered channel measurement samples are frequency-ordered channel measurement samples taken across a set of sub-carrier frequencies. ### Claim 26 (depends on 14) The receiver circuit of claim 14, wherein the receiver circuit is further configured to repeat each of the performing, estimating, computing, and selecting operations for each of a series of channel estimation intervals. ### Claim 27 (depends on 14) A wireless device comprising the receiver circuit of claim 14.