DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 03/20/2026 is in compliance with the provisions of 35 CFR 1.97. Accordingly, the IDS has been considered by the examiner.
Response to Amendment
Applicant's arguments and remarks filed on 01/21/2026 have been fully considered.
Claims 1, 3, 10, 12, and 18 have been amended.
The amendments add to the correcting limitation specific language reciting that the correcting comprises subtracting (Claims 1 and 18) or adjusting (Claims 12 and 18) each measured non-line of sight round trip time based on a respective time difference output by a trained round trip time (RTT) correction machine learning model, the respective time difference corresponding to a difference between the measured non-line of sight round trip time and an expected line of sight round trip time for the measured non-line of sight round trip time, and further add language reciting that the corrected round trip time vector dataset comprises expected line of sight round trip times which are input to a line of sight-based position determination function.
Applicant's amendments overcome the previous U.S.C. 101 rejection.
Applicant's amendments overcome the objections to the specification.
Claims 1–20 are pending.
Response to Arguments
Applicant’s arguments, see remarks pages 10-26, filed 01/21/2026, with respect to the rejection of claims 1-20 under 35 U.S.C. 103 have been fully considered and are persuasive. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground of rejection is made in view of Davis (US 2013/0045755 A1) in view of Tadayon et al. (US 10,908,299 B1).
The rejection of claims 1-20 under 35 U.S.C. 101 is withdrawn. As amended, independent claims 1, 12, and 18 recite a practical application of mathematical concepts involved in correcting round trip time values. The claims now apply trained RTT correction model outputs to measured non-line-of-sight RTTs obtained from wireless communications between transmit-receive points and user equipment to produce expected line-of-sight RTTs and further, use the corrected RTT vector in a line-of-sight based position determination function. From the Examiner point of view, when considered as a whole, the claims reflect a particular solution to a technical problem in RTT-based wireless positioning and integrate the previous judicial exception into a practical application.
Claim Objections
Claims 2, 12, objected to because of the following informalities:
Claim 2 recites “wherein the non-line of sight time measurement data” which now lacks antecedent basis due to the amendment to claim 1 removing “measurement data”. Correct as necessary.
PNG
media_image1.png
135
595
media_image1.png
Greyscale
Claims 12 and 18 recite “trained RT correction machine” which should be “trained RTT correction machine”.
Claim 18 recites “first round trip times comprises” which should be “first round trip times comprise”.
PNG
media_image2.png
164
631
media_image2.png
Greyscale
Appropriate correction is required.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-5, and 8-20 are rejected under 35 U.S.C. 103 as being unpatentable over Davis (US 2013/0045755 A1) in view of Tadayon et al. (US 10,908,299 B1).
Regarding Claim 1, Davis (‘755) in view of Tadayon et al. (‘299) teaches:
Davis (‘755) teaches: A system, comprising: at least one processor; and at least one memory that stores executable instructions that, when executed by the at least one processor, facilitate performance of operations, the operations comprising: ([0049]: “Example system 200 can include a processor 250, which is configured to confer, and confers, at least in part, the described functionality of the various components included in example system 200. Processor 250 can execute code instructions (not shown) stored in memory 240, or other memory(ies), to provide the described functionality.”)
Davis (‘755) teaches: obtaining a dataset comprising round trip time vectors of time measurement data based on communications between a group of transmit-receive points relative to a user equipment at an unknown location, the time measurement data comprising measured non-line of sight round trip times obtained from communications between a transmit-receive point and the user equipment ([0003]: “timing delay of the signals transmitted between the wireless base station and the wireless handset are employed in various location services methods, including, but not limited to, cell global identity and timing advance (CGI+TA), CGI and round trip time (CGI+RTT), time of arrival (TOA), and custom methods. Timing delay is affected by propagation delay in the wireless signal path among radio component(s) at the wireless base station and a sector antenna.”; [0033]: “the network estimates electronically a single overall propagation delay magnitude that is related to an estimate of the time it takes a signal to travel round trip between the base station (BS) radio and the subscriber mobile station (MS) regardless of propagation pathway between BS and MS.”; [0030]: “timing delay spread 188 generally originates from any signal path scattering, or ‘signal bounces,’ such as multipath or strong reflections, etc.”)
Davis (‘755) teaches: correcting the non-line of sight round trip times in the round trip time vector dataset to obtain a corrected round trip time vector dataset comprising expected line of sight time round trip times corresponding to the measured non-line of sight round trip times, wherein the correcting comprises adjusting each measured non- line of sight round trip time of the measured non-line of sight round trip times based on a respective time difference output by a trained round trip time (RTT) correction machine learning model, the respective time difference corresponding to a difference between the measured non-line of sight round trip time and an expected line of sight round trip time for the measured non-line of sight round trip time ([0031]: “a propagation timing delay offset supplies a location correction 198 that compensates the difference between the TOF estimated range 196 and the actual straight-line range 192 from the base station 110.”; [0030]: “calibrated propagation timing substantially reveals LOS timing delay .DELTA..tau..sup.(LOS).”; [0046]: “Analysis component 218 utilizes the timing advance or round trip delay determined by TOF component 212 to correct the timing advance information in a position determination function (PDF) such as CGI+TA.”)
Davis (‘755) does not explicitly teach, but Tadayon et al. (‘299) teaches a respective time difference output by a trained round trip time (RTT) correction machine learning model (Col. 6, lines 41-52: “The error estimates…are then used for corrected positioning determination. For example, the error estimates may be used as labels, with channel data that was collected at each location between a j^th BS and a UE as inputs, to train a Deep Neural Network (DNN) and/or another form of ML module or AI component to be able to generalize to unseen locations.”)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to combine the RTT propagation delay correction system of Davis (‘755) with the trained machine learning model of Tadayon et al. (‘299). One would have been motivated to do so because Davis (‘755) itself recognizes at [0050] that artificial intelligence methods are applicable to determine suitable propagation models (“Various aspects of the subject innovation can be automated through artificial intelligence (AI) methods to infer (e.g., reason and draw a conclusion based upon a set of metrics, arguments, or known outcomes in controlled scenarios) suitable models for propagation of wireless signal, e.g., RF signal, microwave signal, etc.”), and replacing Davis’s iterative statistical correlation approach for offset computation with a trained ML model as taught by Tadayon et al. (‘299) would yield more accurate and generalizable per-link delay offset estimates, particularly in environments with complex and time-varying NLOS multipath conditions that Davis acknowledges are difficult to compensate ([0030]). A person of ordinary skill in the art would have had a reasonable expectation of success because Davis’s reference probe framework already generates exactly the labeled training data pairs - known probe location (yielding expected LOS RTT as distance/speed-of-light) paired with measured TOF/RTT (yielding measured NLOS RTT) and Tadayon et al. (‘299) demonstrates that this can be used to train a DNN to predict NLOS-induced RTT deviations with generalization to unseen locations.
Davis (‘755) teaches: inputting the corrected round trip time vector dataset comprising the expected line of sight round-trip times to a line of sight-based position determination function ([0046]: “Analysis component 218 utilizes the timing advance or round trip delay determined by TOF component 212 to correct the timing advance information in a position determination function (PDF) such as CGI+TA.”; [0022]: “the system(s), method(s), and aspects thereof described herein improve the accuracy of TOF location estimates with respect to estimates provided by conventional systems that ignore, or fail to incorporate, propagation delay in estimation of timing advance (TA) or round trip time (RTT), or any time of flight quantities utilized to estimate location(s) in operational wireless systems.”)
Davis (‘755) teaches: obtaining, in response to the inputting of the corrected round trip time vector dataset, an estimated location of the user equipment ([0021]: “compensated, thus allowing improvement of the accuracy obtained using time of flight (TOF) location estimates, such as Third Generation Partnership Project (3GPP)-defined CGI+TA or CGI+RTT.”; [0031]: “a propagation timing delay offset supplies a location correction 198 that compensates the difference between the TOF estimated range 196 and the actual straight-line range 192 from the base station 110.”)
Regarding Claim 2, Davis (‘755) in view of Tadayon et al. (‘299) teaches the system of claim 1.
Davis (‘755) teaches: wherein the non-line of sight time measurement data obtained from the communications between the transmit-receive point and the user equipment is obtained from first communications between a first transmit-receive point and the user equipment, and wherein the time measurement data further comprises line of sight time measurement data obtained from second communications between a second transmit-receive point and the user equipment ([0025]: “FIG. 1A is a schematic example wireless environment 100 that can operate in accordance with aspects described herein. In particular, example wireless environment 100 illustrates a set of macro cells. Three macro cells 105.sub.1-105.sub.3 comprise the illustrative wireless environment”; [0031]: “FIG. 1C displays an ideal propagation path, e.g., a straight-line pathway, from a location at the base of a tower that supports antenna(s) 112 to a location 192 of mobile 120. The straight-line pathway has an associated straight line TOF timing delay 194.”; [0030]: “timing delay spread 188 generally originates from any signal path scattering, or ‘signal bounces,’ such as multipath or strong reflections, etc.”) Davis (‘755) teaches a multi-cell wireless network with multiple base stations (transmit-receive points), wherein some base station – user equipment communication links experience NLOS signal bounces (first communications from a first transmit-receive point) and other links propagate along straight-line LOS paths (second communications from a second transmit-receive point), as is inherent in the multi-sector wireless environments disclosed throughout.
Regarding Claim 3, Davis (‘755) in view of Tadayon et al. (‘299) teaches the system of claim 1.
Davis (‘755) teaches: wherein the correcting of the non-line of sight round trip times further comprises inputting the time measurement data into the trained RTT correction machine learning model that has been previously trained with round-trip time training data representing round-trip times of a group of communications measured between transmit-receive points of the group of transmit-receive points and device instances at known locations ([0041]: “Additionally or alternatively, in another aspect, location source(s) platform 220 is embodied in one or more network components, such as for example GMLC or SMLC, that supply location estimate(s) 215 embodied in records of position of a set of probes, or wireless beacons, deployed within a coverage cell or sector.”; [0046]: “Calibration platform 210 exploits a first set of known locations of a set of one or more probes (e.g., wireless beacons 520₁–520ₛ) to obtain a second set of location estimates of the known locations using ‘time of flight’ measurements and cell or sector identifiers.”) Davis (‘755) teaches that TOF/RTT measurements from communications between base stations (transmit-receive points) and probes (device instances) at known locations constitute the training data from which the offset is determined.
As noted in the claim 1 analysis, Tadayon et al. (‘299) further teaches the trained ML model using this type of communication-based training data (Col. 6, lines 41-52: “The error estimates may be used as labels, with channel data that was collected at each location between a j.sup.th BS and a UE as inputs, to train a Deep Neural Network (DNN) and/or another form of ML module or AI component to be able to generalize to unseen locations.”)
The motivation to combine Davis (‘755) with Tadayon et al. (‘299) is as stated above with respect to Claim 1.
Regarding Claim 4, Davis (‘755) in view of Tadayon et al. (‘299) teaches the system of claim 3.
Davis (‘755) teaches: wherein the device instances comprise positioning reference units deployed at the known locations ([0041]: “wireless probes can be location equipment, e.g., location measurement units (LMUs), GNSS apparatuses, or probes, deployed in cells or sectors… disparate to a cell or sector that performs the calibration described herein.”; [0041]: “Positions of probes 520.sub.1-520.sub.3 can cover more than one timing advance bands to improve reliability of statistical analysis employed to determine a propagation delay offset… can be stationary or pseudo-stationary”).
Regarding Claim 5, Davis (‘755) in view of Tadayon et al. (‘299) teaches the system of claim 3.
Davis (‘755) teaches: wherein the device instances comprise at least one mobile device configured to report the known locations via global positioning system data ([0038]: “Diagram 400 in FIG. 4A displays a snapshot at an instant .tau.404 of a set 406 of eighteen mobile devices distributed through four TA bands 402.sub.1-402.sub.4 that communicate with a GNSS system 410 (e.g., GPS, Galileo, GLONASS . . . ) through a deep-space link 412. Mobile devices in set 406 receive timing signaling that allows determination, at least in part, of accurate position of each mobile the receives sufficient information (e.g., timing information from three or more satellites) for triangulation.”; [0042]: “In block diagram 420 in FIG. 4B, mobile(s) 430 can be embodied in the set of mobile devices 406. Communication platform 230, which is part of base station 110, receives GNSS-based location estimates 435 over the air-interface (e.g., wireless channel 182) form mobile(s) 430. Location estimate(s) 435, or location information, can be provided at various time instants and aggregated at calibration platform 210 through aggregation component 442 in data management component 211.”) Davis (‘755) teaches that mobile devices receive and deliver accurate location data based on GNSS systems including GPS.
Regarding Claim 8, Davis (‘755) in view of Tadayon et al. (‘299) teaches the system of claim 1.
Davis (‘755) teaches: wherein the transmit-receive points of the group of transmit-receive points are spatially distributed in a deployment environment ([0025]: “FIG. 1A is a schematic example wireless environment 100 that can operate in accordance with aspects described herein. In particular, example wireless environment 100 illustrates a set of macro cells. Three macro cells 105.sub.1-105.sub.3 comprise the illustrative wireless environment; however, it should be appreciated that wireless cellular network deployments can encompass up to 10.sup.4-10.sup.5 coverage macro cells.”; [0028]: “Each macro cell 105.sub..lamda. is sectorized in a 2.pi./3-radians central-angle configuration in which each macro cell includes three sectors, demarcated with dashed lines in FIG. 1.”) Davis (‘755) explicitly teaches multiple base stations serving coverage cells and sectors that are spatially distributed throughout the wireless network deployment environment.
Regarding Claim 9, Davis (‘755) in view of Tadayon et al. (‘299) teaches the system of claim 8.
Davis (‘755) teaches: wherein the transmit-receive points of the group of transmit-receive points are substantially evenly distributed ([0025]: “Three macro cells 105.sub.1-105.sub.3 comprise the illustrative wireless environment; however, it should be appreciated that wireless cellular network deployments can encompass up to 10.sup.4-10.sup.5 coverage macro cells.”; [0028]: “Each macro cell 105.sub..lamda. is sectorized in a 2.pi./3-radians central-angle configuration in which each macro cell includes three sectors, demarcated with dashed lines in FIG. 1”) Davis (‘755) teaches a regular sectorized wireless network architecture with base stations distributed throughout coverage cells in a substantially even cellular grid pattern, each macro cell covering an equivalent angular sector of 2π/3 radians.
Regarding Claim 10, Davis (‘755) in view of Tadayon et al. (‘299) teaches the system of claim 1.
Davis (‘755) teaches: wherein the correcting of the non-line of sight round trip times further comprises inputting the time measurement data into the trained RTT correction machine learning model that has been previously trained via supervised learning with labeled training data associated with the respective transmit-receive points, the labeled training data comprising respective determined line of sight round trip times based on respective locations of respective device instances, and respective measured round trip time data measured via communications between the respective transmit-receive points and the respective device instances at the respective locations ([0046]: “Calibration platform 210 exploits a first set of known locations of a set of one or more probes (e.g., wireless beacons 520.sub.1-520.sub.3) to obtain a second set of location estimates of the known locations using "time of flight" measurements and cell or sector identifiers. Analysis component 218 utilizes the timing advance or round trip delay determined by TOF component 212 to correct the timing advance information in a position determination function (PDF) such as CGI+TA.”; [0048]: “Determined signal path propagation delay offset(s) for a cell or sector can be retained in delay offset storage 246. A delay offset, or delay offset error, compensates for signal path propagation due to one or more of the propagation delay sources described above, and affords to ensure location estimates based upon TOA or TOF methods are accurate.”) Davis (‘755) teaches using the known probe locations (from which determined LOS RTTs are derivable as distance/speed-of-light) paired with measured TOF/RTT communications between base stations and those probes as labeled training data, associated with the respective base station (transmit-receive point) serving each probe.
As noted in the claim 1 analysis, Tadayon et al. (‘299) further teaches the supervised learning methodology explicitly (Col. 6, lines 41-52: “The error estimates may be used as labels, with channel data that was collected at each location between a j.sup.th BS and a UE as inputs, to train a Deep Neural Network (DNN) and/or another form of ML module or AI component to be able to generalize to unseen locations.”)
The motivation to combine Davis (‘755) with Tadayon et al. (‘299) is as stated above with respect to Claim 1.
Regarding Claim 11, Davis (‘755) in view of Tadayon et al. (‘299) teaches the system of claim 10.
Claim 11 recites: wherein the device instances comprise at least one of: a mobile device instance moved among the second known locations, or a positioning reference unit moved among the second known locations. The recited “second known locations” is a contingent element arising from the parent claim limitations. Per MPEP guidance, this element is contingent and need not be separately addressed unless triggered by the parent claim structure. Further, the “or” statement requires only that the art teach a mobile device instance moved among known locations or a positioning reference unit moved among known locations; the art need not teach both.
Davis (‘755) teaches at least one of these alternatives — a positioning reference unit moved among the second known locations ([0041]: “In a further example, wireless beacons can be truck-mounted radio or microwave transceivers, such as those fitted in service or public transportation vehicles, temporarily stationed in disparate locations throughout a sector to facilitate TOF data collection.”) Davis (‘755) teaches truck-mounted wireless beacons (positioning reference units) that are moved among multiple known locations throughout a sector for training data collection.
Regarding Claim 12, Davis (‘755) in view of Tadayon et al. (‘299) teaches:
Davis (‘755) teaches: inputting, by a system comprising at least one processor to a trained round trip time (RTT) correction machine learning model, a dataset comprising round trip time vectors of round trip time data measured via communications between a user equipment at an unknown location and at least some transmit-receive points distributed at first known locations ([0036]: “As described above, TOF component 212 can measure the propagation timing delay via at least in part clock layer 216; the timing delay can include TA, angle of arrival (AOA), RTT, RL Time of Arrival (RL-TOA), RL Time Difference of Arrival (RL-TDOA), FL-TOA, FL-TDOA, or observed TOA (O-TOA)”; [0029]: “Time of flight measurements probe time of arrival, which is the propagation timing, or round trip time which includes propagation timing from the handset to the radio equipment.”) as taught with respect to Claim 1 above.
Davis (‘755) teaches: the trained RT correction machine learning model having been trained via a training process using training round-trip time data between the at least some transmit-receive points and device instances at second known locations, the round-trip time data comprising measured non-line of sight round-trip times corresponding to non-line of sight measurements of at least a portion of the communications ([0046]: “Calibration platform 210 exploits a first set of known locations of a set of one or more probes (e.g., wireless beacons 520.sub.1-520.sub.3) to obtain a second set of location estimates of the known locations using "time of flight" measurements and cell or sector identifiers.”; [0030]: “timing delay spread 188 generally originates from any signal path scattering, or “signal bounces,” such as multipath or strong reflections, etc.”) Davis (‘755) teaches probe devices at second known locations communicating via RTT measurements with base stations, with at least a portion of those communications including NLOS signal bounces and multipath.
Davis (‘755) does not explicitly teach, but Tadayon et al. (‘299) teaches the trained RT correction machine learning model (Col. 6, lines 41-52: “The error estimates may be used as labels, with channel data that was collected at each location between a j.sup.th BS and a UE as inputs, to train a Deep Neural Network (DNN) and/or another form of ML module or AI component to be able to generalize to unseen locations.”)
The motivation to combine Davis (‘755) with Tadayon et al. (‘299) is as stated above with respect to Claim 1.
Davis (‘755) teaches: correcting, by the trained RT correction machine learning model of the system, the measured non-line of sight round-trip times into expected line of sight round-trip times corresponding to the measured non-line of sight round trip times, wherein the correcting comprises adjusting each measured non-line of sight round trip time of the measured non-line of sight round trip times based on respective time difference output by the trained RTT correction machine learning model, the respective time difference corresponding to a difference between the measured non-line of sight round trip time and an expected line of sight round trip time for the measured non-line of sight round trip time ([0031]: “a propagation timing delay offset supplies a location correction 198 that compensates the difference between the TOF estimated range 196 and the actual straight-line range 192 from the base station 110.”; [0030]: “calibrated propagation timing substantially reveals LOS timing delay .DELTA..tau..sup.(LOS).”; [0046]: “Analysis component 218 utilizes the timing advance or round trip delay determined by TOF component 212 to correct the timing advance information in a position determination function (PDF) such as CGI+TA.”) as to the adjusting concept and time difference structure. Tadayon et al. (‘299) teaches the trained RTT correction machine learning model outputting this time difference correction as discussed with respect to Claim 1.
Davis (‘755) teaches: inputting, by the system to a line of sight-based position determination function, a modified round trip time vector dataset comprising the expected line of sight round-trip times; and obtaining, by the system in response to the inputting of the modified round trip time vector dataset, an estimated location of the user equipment ([0046]: “Analysis component 218 utilizes the timing advance or round trip delay determined by TOF component 212 to correct the timing advance information in a position determination function (PDF) such as CGI+TA.”; [0021]: “When an effective total timing delay, which includes propagation delay, is determined, wireless signal propagation delay information can be corrected, or compensated, thus allowing improvement of the accuracy obtained using time of flight (TOF) location estimates, such as Third Generation Partnership Project (3GPP)-defined CGI+TA or CGI+RTT.”) as taught with respect to Claim 1.
Regarding Claim 13, Davis (‘755) in view of Tadayon et al. (‘299) teaches the method of claim 12.
Davis (‘755) teaches: wherein the inputting of the modified round trip time vector dataset further comprises inputting non-corrected line of sight round-trip time data as part of the modified round trip time vector dataset ([0046]: “Analysis component 218 utilizes the timing advance or round trip delay determined by TOF component 212 to correct the timing advance information in a position determination function (PDF) such as CGI+TA.”; [0040]: “Calibration of ETTD and associated timing delay error for range or distance from a base station, and sector bearing or angular position within a served cell or sector can enable refinement of location services based at least in part on CGI+RTT, for example, and improvement of conventional methods that may refine CGI+RTT.”) Davis (‘755) teaches that the position determination function receives corrected RTT data alongside the complete set of base station RTT measurements, which inherently includes non-corrected LOS RTT data from base stations whose links are determined to be LOS, as only NLOS-affected measurements require correction while LOS measurements pass through uncorrected.
Regarding Claim 14, Davis (‘755) in view of Tadayon et al. (‘299) teaches the method of claim 12.
Davis (‘755) teaches: wherein the training process further comprises arranging non-line of sight transmit-receive points between a device of the device instances and the non-line of sight transmit-receive points more densely than line of sight transmit-receive points between the device of the device instances and the line of sight transmit-receive points ([0043]: “Judicious deployment of wireless beacon(s) 560, e.g., high-density deployment of wireless probes in the vicinity of nominal sector boundaries, and sectorization mapping effected by calibration platform 210 can determine actual coverage pattern of one or more sectors in the field.”; [0043]: “deployment of wireless probes in the vicinity of nominal sector boundaries”) Davis (‘755) teaches that probe deployment density and positioning is deliberately tailored to the coverage and multipath characteristics of specific sectors, which includes denser arrangement of transmit-receive points in areas experiencing greater NLOS propagation effects relative to areas with predominantly LOS propagation.
Regarding Claim 15, Davis (‘755) in view of Tadayon et al. (‘299) teaches the method of claim 12.
Davis (‘755) teaches: wherein at least one of the device instances comprises a positioning reference unit, and wherein the training process further comprises moving the positioning reference unit among at least two of the second known locations ([0041]: “In a further example, wireless beacons can be truck-mounted radio or microwave transceivers, such as those fitted in service or public transportation vehicles, temporarily stationed in disparate locations throughout a sector to facilitate TOF data collection.”) Davis (‘755) teaches truck-mounted wireless beacons, serving as positioning reference units, that are temporarily stationed and moved among multiple known locations throughout a sector for training data collection purposes.
Regarding Claim 16, Davis (‘755) in view of Tadayon et al. (‘299) teaches the method of claim 12.
Davis (‘755) teaches: wherein at least one of the device instances comprises a mobile device, and wherein the training process further comprises moving the mobile device among at least two of the second known locations ([0042]: “In example system 540 in FIG. 5B, network management component 550 has access to accurate location records 552 of wireless beacon(s) 560, and delivers location estimate(s) 565 which comprise such records.”; [0038]: “Mobiles in set 406 also can receive assisted timing information from mobile network platform(s) 108 through base station 110 serving sector 405; mobile network platform(s) 108 received timing information from GNSS 410 through deep-space link 414.”) Davis (‘755) teaches mobile devices that communicate RTT measurements with base stations while moving among known GNSS-verified locations throughout the coverage sector, thereby collecting training data at multiple known locations.
Regarding Claim 17, Davis (‘755) in view of Tadayon et al. (‘299) teaches the method of claim 12.
Davis (‘755) teaches: wherein the communications between the user equipment and the at least some transmit-receive points are first communications, wherein the round trip time data is first measured round trip time data, and wherein the training process further comprises obtaining labeled training data comprising respective determined round trip time data based on the second known locations, and second measured round trip time data of second communications, respectively, between the at least some transmit-receive points at the first known locations and the device instances at the second known locations ([0045]: “In example system 200, calibration platform 210 utilizes location estimate(s) 215 and timing delay measurements in at least two manners in order to measure RF signal or microwave signal propagation delay and thus correct the RF signal or microwave signal propagation information. (1) Location estimate(s) 215 obtained from handsets (e.g., mobiles in set 406) that support generation of high accuracy location data are complemented with location estimate(s), of the same ground truth, generated by TOF component 212 as described above, e.g., via CGI+TA or CGI+RTT, or other RL, FL or observed TOA and TDOA approaches”; [0046]: “Calibration platform 210 exploits a first set of known locations of a set of one or more probes (e.g., wireless beacons 520.sub.1-520.sub.3) to obtain a second set of location estimates of the known locations using "time of flight" measurements and cell or sector identifiers.”) Davis (‘755) teaches obtaining labeled training data comprising: (a) determined RTT data derivable from the known probe locations at second known locations relative to the base stations at first known locations (corresponding to “respective determined round trip time data based on the second known locations”), and (b) measured RTT data from second communications between base stations at first known locations and probe device instances at second known locations (corresponding to “second measured round trip time data of second communications, respectively, between the at least some transmit-receive points at the first known locations and the device instances at the second known locations”).
Regarding Claim 18, Davis (‘755) in view of Tadayon et al. (‘299) teaches:
Davis (‘755) teaches: A non-transitory machine-readable medium, comprising executable instructions that, when executed by at least one a processor, facilitate performance of operations, the operations comprising: ([0049]: “Processor 250 can execute code instructions (not shown) stored in memory 240, or other memory(ies), to provide the described functionality.”)
Davis (‘755) teaches: obtaining a vector dataset at a trained round trip time (RTT) correction machine learning model, the vector dataset comprising respective first round trip times measured based on respective first communications between a user equipment at an unknown location and respective first known locations of a first group of respective transmit-receive points, wherein the first round trip times comprises at least one non-line of sight round trip time ([0003]: “timing delay of the signals transmitted between the wireless base station and the wireless handset are employed in various location services methods, including, but not limited to, cell global identity and timing advance (CGI+TA), CGI and round trip time (CGI+RTT).”; [0030]: “timing delay spread 188 generally originates from any signal path scattering, or “signal bounces,” such as multipath or strong reflections, etc.”) as taught with respect to Claim 1 above.
Davis (‘755) teaches: the trained RT correction machine learning model having been trained with labeled training data comprising respective second determined line of sight round trip time training data based on respective second known locations of a second group of the respective transmit-receive points and respective third known locations of training device instances, and respective measured round trip time training data representing measured third respective round trip times of respective training communications between the second group of the respective transmit-receive points and the training device instances, wherein at least one of the respective training communications comprises a non-line of sight communication ([0046]: “Calibration platform 210 exploits a first set of known locations of a set of one or more probes (e.g., wireless beacons 520.sub.1-520.sub.3) to obtain a second set of location estimates of the known locations using "time of flight" measurements and cell or sector identifiers.”; [0030]: “timing delay spread 188 generally originates from any signal path scattering, or “signal bounces,” such as multipath or strong reflections, etc.”; [0048]: “Determined signal path propagation delay offset(s) for a cell or sector can be retained in delay offset storage 246. A delay offset, or delay offset error, compensates for signal path propagation due to one or more of the propagation delay sources described above.”) Davis (‘755) teaches training data comprising: known probe locations (third known locations of training device instances) and corresponding base station locations (second known locations of a second group of transmit-receive points) from which determined LOS RTT training data are derivable, paired with measured RTT training data from communications between those base stations and probe devices including at least one NLOS communication.
As noted in the claim 1 analysis, Tadayon et al. (‘299) further teaches the trained ML model employing this type of labeled training data (Col. 6, lines 41-52: “The error estimates may be used as labels, with channel data that was collected at each location between a j.sup.th BS and a UE as inputs, to train a Deep Neural Network (DNN) and/or another form of ML module or AI component to be able to generalize to unseen locations.”)
Davis (‘755) does not explicitly teach, but Tadayon et al. (‘299) teaches the trained RT correction machine learning model as discussed with respect to Claim 1.
The motivation to combine Davis (‘755) with Tadayon et al. (‘299) is as stated above with respect to Claim 1.
Davis (‘755) teaches: modifying the vector dataset by the model into a modified vector dataset, the modifying of the vector dataset comprising correcting the at least one non-line of sight round trip time into at least one respective expected line of sight round trip time, wherein the correcting comprises adjusting each measured non-line of sight round trip time of the measured non-line of sight round trip times based on respective time difference output by the trained RTT correction machine learning model, the respective time difference corresponding to a difference between the measured non-line of sight round trip time and an expected line of sight round trip time for the measured non-line of sight round trip time ([0031]: “a propagation timing delay offset supplies a location correction 198 that compensates the difference between the TOF estimated range 196 and the actual straight-line range 192 from the base station 110.”; [0030]: “calibrated propagation timing substantially reveals LOS timing delay .DELTA..tau..sup.(LOS).”; [0046]: “Analysis component 218 utilizes the timing advance or round trip delay determined by TOF component 212 to correct the timing advance information in a position determination function (PDF) such as CGI+TA.”) as taught with respect to Claims 1 and 12. Tadayon et al. (‘299) teaches the trained ML model outputting the time difference as discussed with respect to Claim 1.
Davis (‘755) teaches: inputting the modified vector dataset comprising the at least one respective expected line of sight round trip time to a line of sight-based position determination function; and obtaining, in response to the inputting of the modified round trip time vector dataset, an estimated location of the user equipment ([0046]: “Analysis component 218 utilizes the timing advance or round trip delay determined by TOF component 212 to correct the timing advance information in a position determination function (PDF) such as CGI+TA.”; [0021]: “compensated, thus allowing improvement of the accuracy obtained using time of flight (TOF) location estimates, such as Third Generation Partnership Project (3GPP)-defined CGI+TA or CGI+RTT.”) as taught with respect to Claim 1.
Regarding Claim 19, Davis (‘755) in view of Tadayon et al. (‘299) teaches the non-transitory machine-readable medium of claim 18.
Davis (‘755) teaches: wherein the respective first known locations comprise the respective second known locations ([0046]: “Calibration platform 210 exploits a first set of known locations of a set of one or more probes (e.g., wireless beacons 520.sub.1-520.sub.3) to obtain a second set of location estimates of the known locations using "time of flight" measurements and cell or sector identifiers.”) Davis (‘755) teaches that the same base stations at fixed, known positions are used for both operational RTT positioning communications (first known locations for first communications) and as the transmit-receive point reference for training data generation (second known locations for training communications), as the base stations serve both functions within the same calibration and operational framework.
Regarding Claim 20, Davis (‘755) in view of Tadayon et al. (‘299) teaches the non-transitory machine-readable medium of claim 18.
Claim 20 recites: wherein the respective device instances at the third known locations comprise at least one of: a positioning reference unit, or a mobile device. The “or” statement requires only that the art teach a positioning reference unit at the known locations or a mobile device at the known locations.
Davis (‘755) teaches at least one of these alternatives — a positioning reference unit ([0041]: “In yet another example, beacons or wireless probes can be location equipment, e.g., location measurement units (LMUs), GNSS apparatuses, or probes, deployed in cells or sectors, particularly equipment located in cells or sectors disparate to a cell or sector that performs the calibration described herein.”) Davis (‘755) teaches probes and wireless beacons serving as positioning reference units deployed at known third locations as the training device instances.
Claims 6-7 are rejected under 35 U.S.C. 103 as being unpatentable over Davis (US 2013/0045755 A1) in view of Tadayon et al. (US 10,908,299 B1) and further in view of Zhang et al. (US 2016/0323753 A1).
Regarding Claim 6, Davis (‘755) in view of Tadayon et al. (‘299) does not explicitly teach, but Zhang et al. (‘753) further teaches: wherein the transmit-receive points and the device instances at the known locations are represented by a digital twin simulation of an environment, and wherein the round-trip time training data is based on the digital twin simulation (Abstract: “constructing an indoor and outdoor combined three-dimensional scene model of a target building, predicting wireless signal field intensity information of 3D space using a ray-tracing algorithm, selecting a small quantity of testing points to perform manual field measurements and recording the wireless signal intensity information, correcting/calibrating 3D ray-tracing propagation model parameters based on the difference between the actually measured wireless signal intensity information and the wireless signal intensity information calculated through the principle of the 3D ray-tracing propagation model.”)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to further combine the propagation delay correction system of Davis (‘755) and trained ML model of Tadayon et al. (‘299) with the 3D digital twin simulation approach of Zhang et al. (‘753). One would have been motivated to do so in order to generate RTT training data for the ML model from a simulated 3D model of the deployment environment, thereby reducing the cost and logistical burden of physically deploying probes at all training locations throughout a coverage area. A person of ordinary skill in the art would have had a reasonable expectation of success because Zhang et al. (‘753) demonstrates that a 3D ray-tracing model calibrated from a small number of physical measurements can accurately represent wireless signal propagation across an entire environment (Abstract), and Tadayon et al. (‘299) similarly recognizes that simulation-based and crowdsourced data can supplement physical training data collection.
Regarding Claim 7, Davis (‘755) in view of Tadayon et al. (‘299) does not explicitly teach, but Tadayon et al. (‘299) and Zhang et al. (‘753) further teach: wherein the operations further comprise refining spatial resolution of the transmit-receive points via semi-supervised learning
Davis (‘755) does not explicitly teach, but Tadayon et al. (‘299) teaches a semi-supervised approach to training data collection (Col. 4, lines 38-51: “One possible approach to remedy the scalability issue of expert-system fingerprinting or at least mitigate the effects of that issue, in accordance with some embodiments, involves collecting labelled training data through UE-BS communications. This in effect leverages the distributed and mobile nature of UEs in order to collect training datasets and label them using imperfect labelling algorithms.”)
Davis (‘755) and Tadayon et al. (‘299) do not explicitly teach, but Zhang et al. (‘753) teaches using a small quantity of physical measurements to calibrate and refine a broader spatial model (Abstract: “selecting a small quantity of testing points to perform manual field measurements and recording the wireless signal intensity information, correcting/calibrating 3D ray-tracing propagation model parameters based on the difference between the actually measured wireless signal intensity information and the wireless signal intensity information calculated through the principle of the 3D ray-tracing propagation model.”)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to apply semi-supervised learning to refine the spatial resolution of transmit-receive points within the combined Davis (‘755), Tadayon et al. (‘299), and Zhang et al. (‘753) system. One would have been motivated to do so in order to leverage the large volume of unlabeled RTT communications data alongside the limited labeled reference probe data to improve spatial coverage and granularity of the NLOS correction model without requiring exhaustive labeled data collection at every transmit-receive point location. A person of ordinary skill in the art would have had a reasonable expectation of success because semi-supervised learning is a well-established ML technique known to improve model performance by exploiting unlabeled data, and both Tadayon et al. (‘299) and Zhang et al. (‘753) disclose using limited physical measurements to extrapolate correction models across broader spatial environments.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to REMASH R GUYAH whose telephone number is (571)270-0115. The examiner can normally be reached M-F 7:30-4:30.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Resha H Desai can be reached at (571) 270-7792. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/REMASH R GUYAH/Examiner, Art Unit 3648
/RESHA DESAI/Supervisory Patent Examiner, Art Unit 3648
Prosecution Insights