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 .
Response to Amendment
The amendment filed March 30, 2026 has been entered. Claims 1-9, 12-23, and 26-30 remain pending in the application.
Response to Arguments
Rejection of claims 1, 3, 18, and other claims dependent from claim 1, under 35 U.S.C. 112(b) have been withdrawn. However, Applicant failed to amend claim 28 to overcome the rejection under 35 U.S.C. 112(b). Accordingly, rejection of claim 28 under 35 U.S.C. 112(b) is maintained.
Applicant's arguments filed on March 30, 2026 have been fully considered but they are not persuasive.
On pp. 12 and 13 of Applicant’s response, Applicant argues the combination of Tullberg and Sun does not teach the limitations of amended claims 1, 14, 28, and 30. In particular, Applicant argues that “Sun’s ‘process location’ is a data pipeline State [sic], not a geographic location area” and further contends that “[a] reference point in a communication process as taught by Sun has nothing to do with UE positioning.” (Applicant’s response, pp. 12 and 13). Examiner disagrees.
Sun describes that “[p]rocess location information” can be “a node location in a data processing process.” (Sun, ¶ [0043]). Sun further describes “[a] process node included in a physical layer process may be used as a reference point. The physical layer process includes an initial cell search process, a random access procedure, a beam management process, and the like.” (Sun, ¶ [0229]). Since the process node is included in a physical layer process, which includes processes such as an initial cell search process, it inherently discloses that the process node is part of a communications device in the network, and accordingly, the node location does disclose location of the node (i.e., property type compris[ing]: location area information).
Accordingly, the combination of Tullberg and Sun teaches all of the limitations of claims 1, 14, 28, and 30.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 28 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 28 recites “a second set of characteristics associated with a second set of characteristics.” A set of characteristics being associated with itself fails to reasonably apprise one of ordinary skill in the art of the scope of the invention, and thus, renders claim 28 indefinite. For examination purposes, the term “a second set of characteristics associated with a second set of characteristics” has been construed as just a second set of characteristics.
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 1-9, 13-23, and 27-30 are rejected under 35 U.S.C. 103 as being unpatentable over Tullberg et al. (U.S. Publication No. 2022/0078637) in view of Sun et al. (U.S. Publication No. 2024/0396695).
Regarding claim 1, Tullberg teaches “[a] method of operating a communications device, comprising: determining a first quasi-model relation (QML) between a reference machine-learning (ML)-based UE position estimation model and a first ML-based UE position estimation model that is based on a correspondence between a first set of characteristics associated with the reference ML-based UE position estimation model and a second set of characteristics associated with the first ML-based UE position estimation model” (see ¶¶ [0017], [0088], [0089], [0100], [0140], and [0146]; machine learning (ML) models’ output data of prediction operations and parameters include data used for position estimation such as Received Signal Strength Indicator (RSSI), location, and orientation of devices; thus, the described ML models are machine-learning (ML)-based UE position estimation models; furthermore, network node receives information of a prediction of an operation by a first instance of an ML model (i.e., first instance of ML model can be a first ML-based UE position estimation model), and the network node updates one or more parameters of a second instance of the ML model (i.e., second instance of ML model can be a reference machine-learning (ML)-based UE position estimation model); the first instance of the ML model relates to a wireless device and the second instance of the ML model also relates to the wireless device, and to update parameters of the second instance (a reference machine-learning (ML)-based UE position estimation model), the network node must identify/determine that the parameters (i.e., first set of characteristics) of the second instance correspond to the parameters (i.e., second set of characteristics) of the first instance (first ML-based UE position estimation model); and if the parameters (second set of characteristics) of the first instance and the parameters (first set of characteristics) of the second instance correspond to each other, then there exists at least some relation/relationship (quasi-model relation) between the first instance (first ML-based UE position estimation model) and the second instance (reference machine-learning (ML)-based UE position estimation model); thus, a first quasi-model relation (QML) is determined between a reference machine-learning (ML)-based UE position estimation model and a first ML-based UE position estimation model that is based on a correspondence between a first set of characteristics associated with the reference ML-based UE position estimation model and a second set of characteristics); and
As explained above, Tullberg teaches determining a quasi-model relation between the two ML models, and further teaches transmitting indication of whether parameters of first instance of the ML model needs updating. (See ¶¶ [0106] and [0112], and FIG. 2). Transmitting such an indication can implicitly teach transmitting at least some relationship (i.e., quasi-model relation) between the two ML models. But Tullberg does not explicitly disclose “transmitting an indication of the first QML, wherein the first QML indicates a degree of similarity or difference between a first value or first range of values for a property type associated with the first set of characteristics and a second value or second range of values for the property type associated with the first set of characteristics, wherein the property type comprises: location area information, or timing information, or model complexity, or Doppler shift, or Doppler spread, or average delay, or delay spread, or spatial reception filter, or spatial transmission filter, or any combination thereof” of claim 1. However, the foregoing limitations were well known in the art prior to the effective filling date of the claimed invention.
For example, Sun teaches “transmitting an indication of the first QML” (see ¶¶ [0034], [0038], [0039], [0043], [0235], [0236], and [0238]; a communication device obtains a first set of reference points, where one reference point corresponds to an input of a first model and another reference point corresponds to an output of a second model; first and second models can be two of same model or different models and are AI models; a communication device performs model matching based on the reference points (i.e., determining a relation (first QML) between the models), and if the models match, sends (transmits) feedback including acknowledgement (i.e., an indication of the first QML); thus, communication devices transmits an indication of the first QML),
Sun further teaches “wherein the first QML indicates a degree of similarity or difference between a first value or first range of values for a property type associated with the first set of characteristics and a second value or second range of values for the property type associated with the first set of characteristics” (see ¶¶ [0034], [0038], [0039], [0043], [0235], [0236], and [0238]; a communication device performs model matching (i.e., first QML indicating a degree of similarity or difference) based on the received reference points (i.e., a first value or first range of values for a property type associated with the first set of characteristics and a second value or second range of values for the property type associated with the first set of characteristics); thus, the first QML indicates a degree of similarity or difference between a first value or first range of values for a property type associated with the first set of characteristics and a second value or second range of values for the property type associated with the first set of characteristics),
Sun also teaches “wherein the property type comprises: location area information, or timing information, or model complexity, or Doppler shift, or Doppler spread, or average delay, or delay spread, or spatial reception filter, or spatial transmission filter, or any combination thereof” (see ¶¶ [0043] and [0229]; reference points include process location information, which can include location of a node; a process node included in a physical layer process may be used as a reference point; the physical layer process includes an initial cell search process, a random access procedure, a beam management process, and the like; since the process node is included in a physical layer process, which includes processes such as an initial cell search process, it inherently discloses that the process node is part of a communications device in the network, and accordingly, the node location does disclose location of the node (i.e., property type compris[ing]: location area information); thus, the property type comprises: location area information, or timing information, or model complexity, or Doppler shift, or Doppler spread, or average delay, or delay spread, or spatial reception filter, or spatial transmission filter, or any combination thereof).
Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg to incorporate the teachings of Sun to transmit an indication of a relations between two ML models, where the QML indicates a degree of similarity or difference for a property type that is location area information. The suggestion to do so would have been to reduce communication overhead during model interaction/verification (see ¶ [0004] of Sun).
Regarding claim 2, the combination of Tullberg and Sun teaches the method of claim 1, and further teaches “wherein the communications device corresponds to a network component or a UE” (see ¶ [0088] of Tullberg; network node (i.e., a network component) performs the operations; thus, the communications device corresponds to a network component or a UE).
Regarding claim 3, the combination of Tullberg and Sun teaches the method of claim 1, and further teaches “wherein the first ML model and the reference ML model each corresponds to a physical model or a logical model” (see ¶¶ [0064] and [0065] of Tullberg; [NOTE: as explained under the 112(b) rejection, for examination purposes, the foregoing limitation has been construed to correspond to just an ML model]; the machine learning models may correspond to particular site, which is a physical location of a communication device; thus, the first ML model and the reference ML model each corresponds to a physical model).
Regarding claim 4, the combination of Tullberg and Sun teaches the method of claim 1, and further teaches “wherein the first ML-based UE position estimation model and the reference ML-based UE position estimation model are associated with the same location region or different location regions that overlap with each other at least in part” (see ¶¶ [0064] and [0065] of Tullberg; the machine learning models may correspond (i.e., associated with) to particular site, which is a physical location, (i.e., can be same location region or different location regions that overlap with each other at least in part); thus, the first ML-based UE position estimation model and the reference ML-based UE position estimation model are associated with the same location region or different location regions that overlap with each other at least in part).
Regarding claim 5, the combination of Tullberg and Sun teaches the method of claim 1, and further teaches “wherein the reference ML-based UE position estimation model is trained based on first training data and the first ML-based UE position estimation model is trained based on second training data” (see ¶¶ [0137] and [0139] of Tullberg; network node determines a training difference between the first instance (the first ML-based UE position estimation model) and the second instance (the reference ML-based UE position estimation model); based on the training difference, the network node trains the second instance (the reference ML-based UE position estimation model); since there is a training difference and the second instance is trained separately based on this difference, at least the training data used in training the second instance (the reference ML-based UE position estimation model) is different from the training data used in training the first instance (the first ML-based UE position estimation model); thus, the reference ML-based UE position estimation model is trained based on first training data and the first ML-based UE position estimation model is trained based on second training data).
Regarding claim 6, the combination of Tullberg and Sun teaches the method of claim 5, and further teaches “determining a second QML between a second ML-based UE position estimation model and the reference ML-based UE position estimation model that is based on a correspondence between the first set of characteristics and a third set of characteristics associated with the second ML-based UE position estimation model” (see ¶¶ [0056] and [0064] of Tullberg; multiple wireless devices can be served by a network node in a particular site and machine learning models can correspond to a particular site; therefore, another (second) wireless device in the same site can also have another (third) instance (i.e., a second ML-based UE position estimation model) of an ML model, where the second instance, as described above with respect to claims 1 and 5, is the reference ML-based UE position estimation model; from this second wireless device, the network node can receive information of a prediction of an operation by the third instance (second ML-based UE position estimation model), and the network node can similarly update one or more parameters of the second instance of the ML model (reference machine-learning (ML)-based UE position estimation model) based on the received information; and to update parameters of the second instance (a reference machine-learning (ML)-based UE position estimation model), the network node must identify/determine that the parameters (the first set of characteristics) of the second instance correspond to the parameters (i.e., third set of characteristics associated with the second ML-based UE position estimation model) of the third instance (second ML-based UE position estimation model); and if the parameters (third set of characteristics) of the third instance and the parameters (the first set of characteristics) of the second instance correspond to each other, then there exists at least some relation/relationship (second quasi-model relation) between the third instance (second ML-based UE position estimation model) and the second instance (reference machine-learning (ML)-based UE position estimation model); thus, a second QML is determined between a second ML-based UE position estimation model and the reference ML-based UE position estimation model that is based on a correspondence between the first set of characteristics and a third set of characteristics associated with the second ML-based UE position estimation model); and
The combination of Tullberg and Sun also teaches “transmitting the second ML-based UE position estimation model and an indication of the second QML” (see ¶¶ [0034], [0038], [0039], [0235], [0236], and [0238] of Sun; a communication device obtains a first set of reference points, where one reference point corresponds to an input of a first model and another reference point corresponds to an output of a second model; first and second models can be two of same model or different models and are AI models; a communication device performs model matching based on the reference points (i.e., determining a relation (first QML) between the models), and if the models match, sends (transmits) feedback including acknowledgement (i.e., an indication of the second QML) and the matched model (i.e., the second ML-based UE position estimation model) ; thus, communication devices transmits the second ML-based UE position estimation model and an indication of the second QML). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg to incorporate the teachings of Sun to transmit an indication of a relations between two ML models and one of the models. The suggestion to do so would have been to reduce communication overhead during model interaction/verification (see ¶ [0004] of Sun).
Regarding claim 7, the combination of Tullberg and Sun teaches the method of claim 6, and further teaches “wherein the second ML-based UE position estimation model is trained based on third training data” (see ¶¶ [0137] and [0139] of Tullberg; network node can similarly determine a training difference between the third instance (the second ML-based UE position estimation model) and the second instance (the reference ML-based UE position estimation model); based on the training difference, the network node trains the second instance (the reference ML-based UE position estimation model); since there is a training difference and the second instance is trained separately based on this difference, at least the training data used in training the third instance (the second ML-based UE position estimation model) is different from the training data used in training the second instance (the reference ML-based UE position estimation model); thus, the second ML-based UE position estimation model is trained based on third training data).
Regarding claim 8, the combination of Tullberg and Sun teaches the method of claim 5, and further teaches “determining a second QML between a second ML-based UE position estimation model and the first ML-based UE position estimation model that is based on a correspondence between the second set of characteristics and a third set of characteristics associated with the second ML-based UE position estimation model” (see ¶ [0211] of Tullberg; network node can receive ML model errors from multiple wireless devices; thus, from multiple ML models, where one of them can be the first ML-based UE position estimation model and another can be a second QML between a second ML-based UE position estimation model; the network node may first average the model errors before the ML model parameters W are updated; to average the model errors and update their parameters, the network node has to determine the two ML models are related (e.g., same or similar models, and thus a second QML), and that both models include (i.e., correspondence between) the parameters (the second set of characteristics and a third set of characteristics associated with the second ML-based UE position model) that are updated; thus, the network node is determining a second QML between a second ML-based UE position estimation model and the first ML-based UE position estimation model that is based on a correspondence between the second set of characteristics and a third set of characteristics associated with the second ML-based UE position estimation model); and
The combination of Tullberg and Sun also teaches “transmitting the second ML-based UE position estimation model and an indication of the second QML” (see ¶¶ [0034], [0038], [0039], [0235], [0236], and [0238] of Sun; a communication device obtains a first set of reference points, where one reference point corresponds to an input of a first model and another reference point corresponds to an output of a second model; first and second models can be two of same model or different models and are AI models; a communication device performs model matching based on the reference points (i.e., determining a relation (second QML) between the models), and if the models match, sends (transmits) feedback including acknowledgement (i.e., an indication of the second QML) and the matched model (i.e., the second ML-based UE position estimation model); thus, communication devices transmits the second ML-based UE position estimation model and an indication of the second QML). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg to incorporate the teachings of Sun to transmit an indication of a relations between two ML models and one of the models. The suggestion to do so would have been to reduce communication overhead during model interaction/verification (see ¶ [0004] of Sun).
Regarding claim 9, the combination of Tullberg and Sun teaches the method of claim 8, and further teaches “wherein the second ML-based UE position estimation model is trained based on third training data” (see ¶¶ [0137] and [0139] of Tullberg; network node can similarly determine a training difference between the third instance (the second ML-based UE position estimation model) and the second instance (the reference ML-based UE position estimation model); based on the training difference, the network node trains the second instance (the reference ML-based UE position estimation model); since there is a training difference and the second instance is trained separately based on this difference, at least the training data used in training the third instance (the second ML-based UE position estimation model) is different from the training data used in training the second instance (the reference ML-based UE position estimation model); thus, the second ML-based UE position estimation model is trained based on third training data).
Regarding claim 10, the combination of Tullberg and Sun teaches the method of claim 1, and further teaches “wherein the first QML indicates a degree of similarity or difference between a first value or first range of values for a property type associated with the first set of characteristics and a second value or second range of values for the property type associated with the first set of characteristics” (see ¶¶ [0034], [0038], [0039], [0235], [0236], and [0238] of Sun; a communication device performs model matching (i.e., first QML indicating a degree of similarity or difference) based on the received reference points (i.e., a first value or first range of values for a property type associated with the first set of characteristics and a second value or second range of values for the property type associated with the first set of characteristics); thus, the first QML indicates a degree of similarity or difference between a first value or first range of values for a property type associated with the first set of characteristics and a second value or second range of values for the property type associated with the first set of characteristics). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg to incorporate the teachings of Sun where the determined relation indicates a similarity or difference between two ML models. The suggestion to do so would have been to reduce communication overhead during model interaction/verification (see ¶ [0004] of Sun).
Regarding claim 11, the combination of Tullberg and Sun teaches the method of claim 10, and further teaches “wherein the property type comprises: location area information, or timing information, or model complexity, or Doppler shift, or Doppler spread, or average delay, or delay spread, or spatial reception filter, or spatial transmission filter, or any combination thereof” (see ¶ [0043] of Sun; reference points include process location information, which can include location of a node; thus, the property type comprises: location area information, or timing information, or model complexity, or Doppler shift, or Doppler spread, or average delay, or delay spread, or spatial reception filter, or spatial transmission filter, or any combination thereof). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg to incorporate the teachings of Sun to have the property type comprises location area information. The suggestion to do so would have been to reduce communication overhead during model interaction/verification (see ¶ [0004] of Sun).
Regarding claim 13, the combination of Tullberg and Sun teaches the method of claim 1, and further teaches “receiving a request for the first QML, wherein the transmission of the indication of the first QML is in response to the request” (see ¶¶ [0234] and [0238] of Sun; a communication device receives a request for reference points to determine whether a model matches (a request for the first QML) and sends whether the models match in response (transmission of the indication of the first QML is in response to the request)). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg to incorporate the teachings of Sun to have a request for QML and transmit an indication of the QML in response. The suggestion to do so would have been to reduce communication overhead during model interaction/verification (see ¶ [0004] of Sun).
Regarding claim 14, Tullberg teaches “[a] method of operating a position estimation entity, comprising: . . . a first quasi-model relation (QML) between a reference machine-learning (ML)-based user equipment (UE) position estimation model that is associated with a first set of characteristics and a first ML-based UE position estimation model that is associated with a second set of characteristics” (see ¶¶ [0017], [0088], [0089], [0100], [0140], and [0146]; machine learning (ML) models’ output data of prediction operations and parameters include data used for position estimation such as Received Signal Strength Indicator (RSSI), location, and orientation of devices; thus, the described ML models are machine-learning (ML)-based UE position estimation models; furthermore, network node receives information of a prediction of an operation by a first instance of an ML model (i.e., first instance of ML model can be a first ML-based UE position estimation model), and the network node updates one or more parameters of a second instance of the ML model (i.e., second instance of ML model can be a reference machine-learning (ML)-based UE position estimation model); the first instance of the ML model relates to a wireless device and the second instance of the ML model also relates to the wireless device, and to update parameters of the second instance (a reference machine-learning (ML)-based UE position estimation model), the network node must identify/determine that the parameters (i.e., first set of characteristics) of the second instance correspond to the parameters (i.e., second set of characteristics) of the first instance (first ML-based UE position estimation model); and if the parameters (second set of characteristics) of the first instance and the parameters (first set of characteristics) of the second instance correspond to each other, then there exists at least some relation/relationship (quasi-model relation) between the first instance (first ML-based UE position estimation model) and the second instance (reference machine-learning (ML)-based UE position estimation model); thus, a first quasi-model relation (QML) between a reference machine-learning (ML)-based user equipment (UE) position estimation model that is associated with a first set of characteristics and a first ML-based UE position estimation model that is associated with a second set of characteristics); and
Tullberg further teaches “performing one or more actions associated with position estimation of one or more UEs” (see ¶ [0164]; the wireless device updates (one or more actions) one or more parameters of an ML model based on the received information, which causes error of a prediction of an operation by the ML model to zero or small; and output data of prediction operations and parameters include data used for position estimation such as Received Signal Strength Indicator (RSSI), location, and orientation of devices; thus, one or more actions associated with position estimation of one or more UEs based on the indication of the first QML are performed)
Tullberg does not explicitly disclose “receiving an indication of a first quasi-model relation (QML)” and performing actions “based on the indication of the first QML” and “wherein the first QML indicates a degree of similarity or difference between a first value or first range of values for a property type associated with the first set of characteristics and a second value or second range of values for the property type associated with the first set of characteristics, wherein the property type comprises: location area information, or timing information, or model complexity, or Doppler shift, or Doppler spread, or average delay, or delay spread, or spatial reception filter, or spatial transmission filter, or any combination thereof” of claim 14. However, the foregoing limitations were well known in the art prior to the effective filling date of the claimed invention. For example, Sun teaches “receiving an indication of a first quasi-model relation (QML)” (see ¶¶ [0034], [0038], [0039], [0235], [0236], and [0238]; the B-end communication device obtains a first set of reference points, where one reference point corresponds to an input of a first model and another reference point corresponds to an output of a second model; first and second models can be two of same model or different models and are AI models; a communication device performs model matching based on the reference points (i.e., determining a relation (first QML) between the models), and if the models match, the A-end communication device receives feedback including acknowledgement (i.e., an indication of the first QML); thus, communication device receives an indication of the first QML), and
Sun also teaches performing actions “based on the indication of the first QML” (see ¶¶ [0043], [0238], and [0239]; the A-end communication device receives uses the mode receives feedback (the indication of the first QML) and uses the matched model (i.e., performs actions) based on the received feedback (based on the indication of the first QML)).
Sun further teaches “wherein the first QML indicates a degree of similarity or difference between a first value or first range of values for a property type associated with the first set of characteristics and a second value or second range of values for the property type associated with the first set of characteristics” (see ¶¶ [0034], [0038], [0039], [0043], [0235], [0236], and [0238]; a communication device performs model matching (i.e., first QML indicating a degree of similarity or difference) based on the received reference points (i.e., a first value or first range of values for a property type associated with the first set of characteristics and a second value or second range of values for the property type associated with the first set of characteristics); thus, the first QML indicates a degree of similarity or difference between a first value or first range of values for a property type associated with the first set of characteristics and a second value or second range of values for the property type associated with the first set of characteristics),
Sun also teaches “wherein the property type comprises: location area information, or timing information, or model complexity, or Doppler shift, or Doppler spread, or average delay, or delay spread, or spatial reception filter, or spatial transmission filter, or any combination thereof” (see ¶¶ [0043] and [0229]; reference points include process location information, which can include location of a node; a process node included in a physical layer process may be used as a reference point; the physical layer process includes an initial cell search process, a random access procedure, a beam management process, and the like; since the process node is included in a physical layer process, which includes processes such as an initial cell search process, it inherently discloses that the process node is part of a communications device in the network, and accordingly, the node location does disclose location of the node (i.e., property type compris[ing]: location area information); thus, the property type comprises: location area information, or timing information, or model complexity, or Doppler shift, or Doppler spread, or average delay, or delay spread, or spatial reception filter, or spatial transmission filter, or any combination thereof).
Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg to incorporate the teachings of Sun to transmit an indication of a relations between two ML models, where the QML indicates a degree of similarity or difference for a property type that is location area information. The suggestion to do so would have been to reduce communication overhead during model interaction/verification (see ¶ [0004] of Sun).
Regarding claim 15, the combination of Tullberg and Sun teaches the method of claim 14, and further teaches “wherein the reference ML-based UE position estimation model is trained based on first training data and the first ML-based UE position estimation model is trained based on second training data” (see ¶¶ [0137] and [0139] of Tullberg; network node determines a training difference between the first instance (the first ML-based UE position estimation model) and the second instance (the reference ML-based UE position estimation model); based on the training difference, the network node trains the second instance (the reference ML-based UE position estimation model); since there is a training difference and the second instance is trained separately based on this difference, at least the training data used in training the second instance (the reference ML-based UE position estimation model) is different from the training data used in training the first instance (the first ML-based UE position estimation model); thus, the reference ML-based UE position estimation model is trained based on first training data and the first ML-based UE position estimation model is trained based on second training data).
Regarding claim 16, the combination of Tullberg and Sun teaches the method of claim 14, and further teaches “wherein the one or more actions comprise activation, deactivation, or selection of the first ML-based UE position estimation model, the reference ML-based UE position estimation model, or another ML-based UE position estimation model, or wherein the one or more actions comprise switching between ML-based UE position estimation models, or a combination thereof” (see ¶ [0164] of Tullberg and ¶ [0239] of Sun; Tullberg teaches the wireless device updates (one or more actions) one or more parameters of an ML model (i.e., activation of the first ML-based UE position estimation model) based on the received information; Sun teaches the A-end communication device uses (one or more actions) the matched model (i.e., activation of the first ML-based UE position estimation model)). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg to incorporate the teachings of Sun to activate the ML model by updating the model’s parameters and using the model. The suggestion to do so would have been to reduce communication overhead during model interaction/verification (see ¶ [0004] of Sun).
Regarding claim 17, the combination of Tullberg and Sun teaches the method of claim 14, and further teaches “wherein the position estimation entity corresponds to a network component or a UE” (see ¶ [0088] of Tullberg; output of prediction of an operation (position estimation operation) of a ML model in wireless device (UE); thus, the position estimation entity corresponds to a network component or a UE).
Regarding claim 18, the combination of Tullberg and Sun teaches the method of claim 14, and further teaches “wherein the first ML model and the reference ML model each corresponds to a physical model or a logical model” (see ¶¶ [0064] and [0065] of Tullberg; [NOTE: as explained under the 112(b) rejection, for examination purposes, the foregoing limitation has been construed to correspond to just an ML model]; the machine learning models may correspond to particular site, which is a physical location of a communication device; thus, the first ML model and the reference ML model each corresponds to a physical model).
Regarding claim 19, the combination of Tullberg and Sun teaches the method of claim 14, and further teaches “wherein the first ML-based UE position estimation model and the reference ML-based UE position estimation model are associated with the same location region or different location regions that overlap with each other at least in part” (see ¶¶ [0064] and [0065] of Tullberg; the machine learning models may correspond (i.e., associated with) to particular site, which is a physical location, (i.e., can be same location region or different location regions that overlap with each other at least in part); thus, the first ML-based UE position estimation model and the reference ML-based UE position estimation model are associated with the same location region or different location regions that overlap with each other at least in part).
Regarding claim 20, the combination of Tullberg and Sun teaches the method of claim 14, and further teaches “receiving an indication of a second QML between a second ML-based UE position estimation model that is associated with a third set of characteristics and the reference ML-based UE position estimation model that is based on a correspondence between the first set of characteristics and a third set of characteristics associated with the second ML-based UE position estimation model” (see ¶¶ [0056] and [0064] of Tullberg, and ¶¶ [0034], [0038], [0039], [0235], [0236], and [0238] of Sun; Tullberg teaches multiple wireless devices can be served by a network node in a particular site and machine learning models can correspond to a particular site; therefore, another (second) wireless device in the same site can also have another (third) instance (i.e., a second ML-based UE position estimation model) of an ML model, where the second instance, as described above with respect to claims 1 and 5, is the reference ML-based UE position estimation model; from this second wireless device, the network node can receive information of a prediction of an operation by the third instance (second ML-based UE position estimation model), and the network node can similarly update one or more parameters of the second instance of the ML model (reference machine-learning (ML)-based UE position estimation model) based on the received information; and to update parameters of the second instance (a reference machine-learning (ML)-based UE position estimation model), the network node must identify/determine that the parameters (the first set of characteristics) of the second instance correspond to the parameters (i.e., third set of characteristics associated with the second ML-based UE position estimation model) of the third instance (second ML-based UE position estimation model); and if the parameters (third set of characteristics) of the third instance and the parameters (the first set of characteristics) of the second instance correspond to each other, then there exists at least some relation/relationship (second quasi-model relation) between the third instance (second ML-based UE position estimation model) and the second instance (reference machine-learning (ML)-based UE position estimation model); Sun teaches the B-end communication device obtains a first set of reference points, where one reference point corresponds to an input of a first model and another reference point corresponds to an output of a second model; first and second models can be two of same model or different models and are AI models; a communication device performs model matching based on the reference points (i.e., determining a relation (second QML) between the models), and if the models match, the A-end communication device receives feedback including acknowledgement (i.e., an indication of the second QML); thus, communication device receives an indication of the second QML),
the combination of Tullberg and Sun further teaches “wherein the one or more actions are further based on the indication of the second QML” (see ¶ [0164] of Tullberg and ¶¶ [0043], [0238], and [0239] of Sun; Tullberg teaches the wireless device updates (one or more actions) one or more parameters of an ML model based on the received information, which causes error of a prediction of an operation by the ML model to zero or small; and output data of prediction operations and parameters include data used for position estimation such as Received Signal Strength Indicator (RSSI), location, and orientation of devices; Sun teaches the A-end communication device receives uses the mode receives feedback (the indication of the first QML) and uses the matched model (i.e., performs actions) based on the received feedback (based on the indication of the second QML)).
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg to incorporate the teachings of Sun to receive an indication of a second relation between two ML models and perform one or more operations associated with position estimation based on the indication of a second relation between the two ML models. The suggestion to do so would have been to reduce communication overhead during model interaction/verification (see ¶ [0004] of Sun).
Regarding claim 21, the combination of Tullberg and Sun teaches the method of claim 20, and further teaches “wherein the second ML-based UE position estimation model is trained based on third training data” (see ¶¶ [0137] and [0139] of Tullberg; network node can similarly determine a training difference between the third instance (the second ML-based UE position estimation model) and the second instance (the reference ML-based UE position estimation model); based on the training difference, the network node trains the second instance (the reference ML-based UE position estimation model); since there is a training difference and the second instance is trained separately based on this difference, at least the training data used in training the third instance (the second ML-based UE position estimation model) is different from the training data used in training the second instance (the reference ML-based UE position estimation model); thus, the second ML-based UE position estimation model is trained based on third training data).
Regarding claim 22, the combination of Tullberg and Sun teaches the method of claim 14, and further teaches “receiving an indication of a second QML between a second ML-based UE position estimation model that is associated with a third set of characteristics and the first ML-based UE position estimation model that is based on a correspondence between the second set of characteristics and a third set of characteristics associated with the second ML-based UE position estimation model” (see ¶ [0211] of Tullberg and ¶¶ [0034], [0038], [0039], [0235], [0236], and [0238] of Sun; Tullberg teaches network node can receive ML model errors from multiple wireless devices; thus, from multiple ML models, where one of them can be the first ML-based UE position estimation model and another can be a second QML between a second ML-based UE position estimation model; the network node may first average the model errors before the ML model parameters W are updated; to average the model errors and update their parameters, the network node has to determine the two ML models are related (e.g., same or similar models, and thus a second QML), and that both models include (i.e., correspondence between) the parameters (the second set of characteristics and a third set of characteristics associated with the second ML-based UE position model) that are updated; Sun teaches the B-end communication device obtains a first set of reference points, where one reference point corresponds to an input of a first model and another reference point corresponds to an output of a second model; first and second models can be two of same model or different models and are AI models; a communication device performs model matching based on the reference points (i.e., determining a relation (second QML) between the models), and if the models match, the A-end communication device receives feedback including acknowledgement (i.e., an indication of the second QML); thus, communication device receives an indication of the second QML),
the combination of Tullberg and Sun further teaches “wherein the one or more actions are further based on the indication of the second QML” (see ¶ [0164] of Tullberg and ¶¶ [0043], [0238], and [0239] of Sun; Tullberg teaches the wireless device updates (one or more actions) one or more parameters of an ML model based on the received information, which causes error of a prediction of an operation by the ML model to zero or small; and output data of prediction operations and parameters include data used for position estimation such as Received Signal Strength Indicator (RSSI), location, and orientation of devices; Sun teaches the A-end communication device receives uses the mode receives feedback (the indication of the first QML) and uses the matched model (i.e., performs actions) based on the received feedback (based on the indication of the second QML)).
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg to incorporate the teachings of Sun to receive an indication of a second relation between two ML models and perform one or more operations associated with position estimation based on the indication of a second relation between the two ML models. The suggestion to do so would have been to reduce communication overhead during model interaction/verification (see ¶ [0004] of Sun).
Regarding claim 23, the combination of Tullberg and Sun teaches the method of claim 22, and further teaches “wherein the second ML-based UE position estimation model is trained based on third training data” (see ¶¶ [0137] and [0139] of Tullberg; network node can similarly determine a training difference between the third instance (the second ML-based UE position estimation model) and the second instance (the reference ML-based UE position estimation model); based on the training difference, the network node trains the second instance (the reference ML-based UE position estimation model); since there is a training difference and the second instance is trained separately based on this difference, at least the training data used in training the third instance (the second ML-based UE position estimation model) is different from the training data used in training the second instance (the reference ML-based UE position estimation model); thus, the second ML-based UE position estimation model is trained based on third training data).
Regarding claim 24, the combination of Tullberg and Sun teaches the method of claim 14, and further teaches “wherein the first QML indicates a degree of similarity or difference between a first value or first range of values for a property type associated with the first set of characteristics and a second value or second range of values for the property type associated with the first set of characteristics” (see ¶¶ [0034], [0038], [0039], [0235], [0236], and [0238] of Sun; a communication device performs model matching (i.e., first QML indicating a degree of similarity or difference) based on the received reference points (i.e., a first value or first range of values for a property type associated with the first set of characteristics and a second value or second range of values for the property type associated with the first set of characteristics); thus, the first QML indicates a degree of similarity or difference between a first value or first range of values for a property type associated with the first set of characteristics and a second value or second range of values for the property type associated with the first set of characteristics). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg to incorporate the teachings of Sun where the determined relation indicates a similarity or difference between two ML models. The suggestion to do so would have been to reduce communication overhead during model interaction/verification (see ¶ [0004] of Sun).
Regarding claim 25, the combination of Tullberg and Sun teaches the method of claim 14, and further teaches “wherein the property type comprises: location area information, or timing information, or model complexity, or Doppler shift, or Doppler spread, or average delay, or delay spread, or spatial reception filter, or spatial transmission filter, or any combination thereof” (see ¶ [0043] of Sun; reference points include process location information, which can include location of a node; thus, the property type comprises: location area information, or timing information, or model complexity, or Doppler shift, or Doppler spread, or average delay, or delay spread, or spatial reception filter, or spatial transmission filter, or any combination thereof). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg to incorporate the teachings of Sun to have the property type comprises location area information. The suggestion to do so would have been to reduce communication overhead during model interaction/verification (see ¶ [0004] of Sun).
Regarding claim 27, the combination of Tullberg and Sun teaches the method of claim 14, and further teaches “transmitting a request for the first QML, wherein the reception of the indication of the first QML is in response to the request” (see ¶¶ [0234] and [0238] of Sun; a communication device sends (transmitting) a request for reference points to determine whether a model matches (a request for the first QML) and receives whether the models match in response (reception of the indication of the first QML is in response to the request)). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg to incorporate the teachings of Sun to have a request for QML and receive an indication of the QML in response. The suggestion to do so would have been to reduce communication overhead during model interaction/verification (see ¶ [0004] of Sun).
Regarding claims 28 and 29, they are apparatus claims corresponding to claims 1 and 2, respectively, that have been rejected above. Applicant’s attention is directed to the rejection of claims 1 and 2. Claims 28 and 29 are rejected under the same rationale.
Regarding claim 30, it is an apparatus claim corresponding to claim 14 that has been rejected above. Applicant’s attention is directed to the rejection of claim 14. Claim 30 is rejected under the same rationale.
Claims 12 and 26 are rejected under 35 U.S.C. 103 as being unpatentable over Tullberg in view of Sun and further in view of Hasegawa et al. (U.S. Publication No. 2024/0295625).
Regarding claim 12, the combination of Tullberg and Sun teaches the method of claim 1, and teaches “the transmission of the indication of the first QML” of claim 12 (see ¶ [0238] of Sun). The combination of Tullberg and Sun does not explicitly disclose transmission “is performed via capability exchange procedure, or wherein the transmission of the indication of the first QML is performed via long term evolution (LTE) positioning protocol (LPP) location request signaling, or wherein the transmission of the indication of the first QML is performed via LPP assistance data signaling, or wherein the transmission of the indication of the first QML is performed via LPP broadcast positioning signaling, or any combination thereof” of claim 12. However, the foregoing limitations were well known in the art prior to the effective filling date of the claimed invention.
For example, Hasegawa teaches “wherein the transmission of the indication of the first QML is performed via capability exchange procedure, or wherein the transmission of the indication of the first QML is performed via long term evolution (LTE) positioning protocol (LPP) location request signaling, or wherein the transmission of the indication of the first QML is performed via LPP assistance data signaling, or wherein the transmission of the indication of the first QML is performed via LPP broadcast positioning signaling, or any combination thereof” (see ¶¶ [0097] and [0165]; communication device sends data via LTE Positioning Protocol (LPP); thus, transmission of data via LPP broadcast positioning signaling). Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg in view of Sun to incorporate the teachings of Hasegawa to have transmit indication of QML via LTE positioning protocol (LPP) broadcast positioning signaling. The suggestion to do so would have been to improve positioning (see ¶ [0012] of Hasegawa).
Regarding claim 26, the combination of Tullberg and Sun teaches the method of claim 1, and further teaches ““the reception of the indication of the first QML” of claim 12 (see ¶ [0238] of Sun). The combination of Tullberg and Sun does not explicitly disclose “wherein the reception of the indication of the first QML is performed via capability exchange procedure, or wherein the reception of the indication of the first QML is performed via long term evolution (LTE) positioning protocol (LPP) location request signaling, or wherein the reception of the indication of the first QML is performed via LPP assistance data signaling, or wherein the reception of the indication of the first QML is performed via LPP broadcast positioning signaling, or any combination thereof” of claim 12. However, the foregoing limitations were well known in the art prior to the effective filling date of the claimed invention.
For example, Hasegawa teaches “wherein the transmission of the indication of the first QML is performed via capability exchange procedure, or wherein the transmission of the indication of the first QML is performed via long term evolution (LTE) positioning protocol (LPP) location request signaling, or wherein the transmission of the indication of the first QML is performed via LPP assistance data signaling, or wherein the transmission of the indication of the first QML is performed via LPP broadcast positioning signaling, or any combination thereof” (see ¶¶ [0097] and [0165]; communication device recieves data via LTE Positioning Protocol (LPP); thus, transmission of data via LPP broadcast positioning signaling). Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the invention of Tullberg in view of Sun to incorporate the teachings of Hasegawa to have transmit indication of QML via LTE positioning protocol (LPP) broadcast positioning signaling. The suggestion to do so would have been to improve positioning (see ¶ [0012] of Hasegawa).
Conclusion
THIS ACTION IS MADE FINAL. 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 SRIHARSHA REDDY VANGAPATY whose telephone number is (571)272-7655. The examiner can normally be reached M-F 8-5 EST.
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, Khaled Kassim can be reached at (571) 270-3770. 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.
/SRIHARSHA REDDY VANGAPATY/ Examiner, Art Unit 2475
/HASHIM S BHATTI/ Primary Examiner, Art Unit 2475
Prosecution Insights