Method and apparatus for generating a predictive model

Abstract

A method and apparatus for generating a predictive model includes operating a wireless signal receiving device while collecting data corresponding to at least one frequency-related parameter associated with a frequency source associated with the receiving device, at least one operating characteristic of the receiving device, and / or at least one environmental event associated with the receiving device, wherein the at least one frequency-related parameter changes with changes in the at least one operating characteristic and / or the at least one environmental event; and processing the data to generate a predictive model used to compensate for changes in the at least one frequency-related parameter corresponding to the at least one operating characteristic and / or the at least one environmental event.

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of and priority to U.S. Provisional Patent Application No. 63 / 445,757, filed February 15, 2023, which is incorporated herein by reference in its entirety.

[0002] FIELD Embodiments of the present principles relate generally to wireless signal processing, and more particularly to methods and apparatus for generating predictive models that compensate for frequency-related parameter errors. [Background technology]

[0003] Radio transmissions are used in various communication and positioning systems. For example, in a handset (e.g., a mobile phone), a frequency source (e.g., an oscillator, a synthesizer, etc.) generates an oscillating signal that is used by various components of the handset to process data and facilitate the operation of various signal receivers (e.g., a WiFi modem, a cellular modem, a Bluetooth modem, a GNSS receiver, etc.). Unfortunately, in most handsets, the stability of the oscillating signal varies greatly depending on the environmental and operating conditions of the handset. Such frequency variations affect the receiver's ability to accurately process received radio signals.

[0004] Techniques for performing motion compensation of received radio signals to improve the accuracy of local frequency sources are known as SUPERCORRELATION™ and are described in commonly assigned U.S. Patent No. 9,780,829, issued October 3, 2017; U.S. Patent No. 10,321,430, issued June 11, 2019; U.S. Patent No. 10,816,672, issued October 27, 2020; U.S. Patent Publication No. 2020 / 0264317, issued August 20, 2020; U.S. Patent Publication No. 2020 / 0319347, issued October 8, 2020; U.S. Patent Application No. 63 / 394667, filed August 3, 2022; and U.S. Patent Application No. 63 / 424185, filed November 10, 2022. The SUPERCORRELATION™ technique can be used to phase compensate the received signal for handset motion and local frequency source instability. In performing the SUPERCORRELATION™ technique, the handset processor generates several phase compensation hypotheses that can correct the received signal for local frequency source instability as well as receiver motion.

[0005] Initially, the hypotheses cover a large search space; over time, as the frequency source stabilizes, the number and density of hypotheses are reduced, and the hypotheses track and correct for changes in the frequency source output. Unfortunately, the output of frequency sources, especially inexpensive oscillators, can fluctuate significantly given the operating characteristics of the handset and / or environmental events. For example, external and internal temperature changes and / or physical shocks to the handset can cause significant and rapid changes in the frequency of the frequency source output. Such unexpected variations require SUPERCORRELATION™ techniques, which expand the hypothesis search space to find a preferred hypothesis that, for example, maximizes the receiver signal correlation output. Such search space expansion can slow signal acquisition, increase receiver power consumption, or make position location using GNSS signals inaccurate for a period of time.

[0006] Therefore, there is a need for a method and apparatus for generating predictive models that anticipate frequency changes resulting from environmental events and device operating characteristics. Summary of the Invention

[0007] SUMMARY OF THE INVENTION Embodiments of the present principles generally relate to methods and apparatus for generating predictive models to anticipate frequency changes resulting from environmental events and / or device operating characteristics.

[0008] The various features and advantages of the present disclosure may be understood by consideration of the following detailed description of the disclosure when taken in conjunction with the accompanying drawings, in which like reference numerals refer to like parts throughout. [Brief explanation of the drawings]

[0009] So that the various features of the present principles can be understood in detail, a more particular description of the principles briefly summarized above can be had by reference to embodiments, some of which are illustrated in the accompanying drawings. It should be noted, however, that the accompanying drawings illustrate only typical embodiments in accordance with the present principles and therefore should not be considered as limiting the scope thereof, as the present principles may admit of other equally effective embodiments.

[0010] [Figure 1] 1 depicts a high-level block diagram of an exemplary testing platform for generating predictive models, in accordance with at least one embodiment of the present principles; [Figure 2] 1 depicts a flow diagram of a method for generating a dataset of test data, in accordance with at least one embodiment of the present principles; [Figure 3] 3 depicts a flow diagram of a method for using the dataset of FIG. 2 to generate a predictive model, in accordance with at least one alternative embodiment of the present principles.

[0011] For ease of understanding, like reference numerals have been used, where possible, to designate like elements common to the figures. The figures are not drawn to scale and may be simplified for clarity. It is contemplated that elements and features of one embodiment may be beneficially incorporated in other embodiments without further description. DETAILED DESCRIPTION OF THE INVENTION

[0012] Embodiments of the present principles generally relate to methods and apparatus for generating predictive models to anticipate frequency changes resulting from environmental events and / or device operating characteristics. While the concepts of the present principles are susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and are described in detail below. It is to be understood that there is no intention to limit the concepts of the present principles to the particular forms disclosed. On the contrary, the intention is to cover all modifications, equivalents, and alternatives consistent with the scope of the present principles and the appended claims.

[0013] As used herein, the phrase device under test (DUT) is intended to describe a receiving device, such as a wireless signal receiving device, that may implement embodiments of the present principles to generate predictive models for anticipating frequency changes resulting from environmental events and / or device operating characteristics, and to implement such predictive models for correcting the frequency response of a frequency source associated with the DUT to changes in the DUT's environmental events and / or device operating characteristics. The phrase device under test is not intended to limit any receiving device to collecting data during a test mode, as embodiments of the present principles described herein include collecting data during normal operation of the receiving device.

[0014] In embodiments of the present principles, the generated predictive model can be used to compensate for expected frequency-related parameter changes that occur during signal processing within a device (i.e., user equipment) in response to device operating characteristics and / or environmental events. Frequency-related parameters can include, but are not limited to, the phase, frequency, frequency rate, or higher-order terms of the device's frequency source. Variations in frequency-related parameters due to environmental and operating conditions can be modeled through a priori empirical studies and / or through real-time updates of the model as the device is used. In some embodiments, the model of the present principles can include a look-up table or machine learning model that associates environmental or operational events with changes in frequency-related parameters. In some embodiments of the present principles, changes in frequency-related parameters can be represented by one or more phasors (i.e., signal phase hypotheses), used, for example, in SUPERCORRELATION™ techniques, to compensate for the predicted frequency-related parameter variations.

[0015] In some embodiments, one or more models can be input using a priori empirical data generated, for example, through handset testing using a test platform. For example, embodiments of the present principles can be used to create a dataset of frequency change information as a handset (e.g., a device under test (DUT)) experiences various environmental conditions (e.g., temperature changes, physical shock, vibration) and / or operational events (e.g., activation and deactivation of a screen, modem, GNSS receiver, etc.). SUPERCORRELATION™ techniques can be used to accurately measure the variations in frequency-related parameters. The dataset can then be processed to create a predictive model that anticipates changes in specific frequency-related parameters that will occur when specific environmental or operational events occur. Models of the present principles allow the number and density of hypotheses used in the SUPERCORRELATION™ techniques to be constrained, for example, when receiving signals during normal use. In embodiments of the present principles, a dataset of information including conditions / events versus frequency-related parameters can be created. Such a dataset may include information describing changes in at least one frequency-related parameter (phase, frequency, frequency rate, or higher-order terms) that occur given a combination of conditions / events (e.g., the frequency and frequency rate changes resulting from dropping a handset on a hot day). In some embodiments, a model created from the dataset may be in the form of a look-up table, or the dataset may be used to train a machine learning model. In some embodiments, a model of the present principles may be used to predict a frequency-related parameter that occurs as a result of the occurrence of a condition / event or combination of conditions / events.

[0016] In some embodiments, a model of the present principles can be initially created using a priori empirical information that is implanted into a new device as an initial predictive model. Such an initial predictive model of the present principles can be updated and / or corrected as the device is used over time. As a result, as the device is used, the predictive model becomes more accurate in predicting frequency variations as the device is used.

[0017] Alternatively, in some embodiments, models of the present principles can be generated in real time as a device (e.g., a handset / mobile communications device) is used, without the implementation of an a priori created dataset. For example, in such embodiments, a predictive model generator of the present principles can be embedded in the handset rather than as part of a testing platform.

[0018] In other embodiments, models of the present principles and / or information about the models can be shared between devices so that the accuracy of the models and / or device functionality is improved. In such embodiments, sharing can be accomplished on a peer-to-peer basis and / or via a server that distributes available model information. For particular models of devices, the server can include reference models developed from the shared information that can be shared periodically with devices currently in use and / or used as initial models on new or reset devices.

[0019] Figure 1 illustrates a block diagram of an exemplary test platform 100 for creating a frequency source predictive model in accordance with at least one embodiment of the present principles. In the embodiment of Figure 1, platform 100 includes a device under test (DUT) 102 coupled to a predictive model generator 128 and an environmental event controller 126. Predictive model generator 128 executes test scripts to collect empirical data from DUT 102 regarding frequency variations in response to specific environmental and operational events. In some embodiments of the present principles, platform 100 of Figure 1 may be implemented, for example, at a DUT manufacturing facility.

[0020] In some embodiments, the predictive model generator 128 may be embedded within the DUT 102 and may run as a software application on the DUT 102. In such embodiments, the generator 128 may function to update the predictive model as the DUT 102 is used in the field.

[0021] In some embodiments, the DUT 102 may include, but is not limited to, a laptop computer, a mobile phone, a tablet computer, an Internet of Things (IoT) device, a dedicated positioning device, etc., and / or combinations thereof. In general, the DUT 102 of FIG. 1 may include any device capable of implementing SUPERCORRELATION™ techniques to enhance signal reception and / or alignment of local frequency sources.

[0022] 1 , DUT 102 illustratively includes at least one processor 104, support circuits 108, and memory 110. At least one processor 104 may include any form of processor or combination of processors, including, but not limited to, a central processing unit, a microprocessor, a microcontroller, a field programmable gate array, a graphics processing unit, a digital signal processor, etc. Support circuits 108 may include well-known circuits and devices that facilitate the function of the processor. That is, support circuits 108 may include one or more and / or combinations of power supplies, clock circuits, analog-to-digital converters, communication circuits, cache, a display, etc.

[0023] In the embodiment of FIG. 1, operation and / or changes in operation of the GNSS receiver 112, the graphical user interface display 114, and the modem 116 (e.g., receiver and transmitter of WiFi, cellular, Bluetooth, etc. signals) can cause temperature changes in the DUT 102, which can result in frequency drift of the frequency source 118.

[0024] 1, memory 110 may include one or more forms of non-transitory computer-readable media, including one or more of read-only memory or random-access memory, or any combination thereof. Memory 110 may store software and data, including, but not limited to, an operating system (OS) 120, application software 122, test software 124 and data 106, signal processing software 146, and SUPERCORRELATION™ software 148, for example.

[0025] In the embodiment of FIG. 1, the predictive model generator 128 includes at least one processor 130, support circuits 132, and memory 134. The at least one processor 130 may include any form of processor or combination of processors, including, but not limited to, a central processing unit, a microprocessor, a microcontroller, a field programmable gate array, a graphics processing unit, a digital signal processor, etc. The support circuits 132 of FIG. 1 may include well-known circuits and devices that facilitate the function of the processor. In the embodiment of FIG. 1, the support circuits 108 may include one or more, or combinations of, power supplies, clock circuits, analog-to-digital converters, communication circuits, cache, a display, test sensors, DUT interface and control ports and circuits, environmental event controller interfaces, and / or the like.

[0026] 1 may include one or more forms of non-transitory computer-readable media, including one or more of read-only memory or random-access memory, or any combination thereof. Memory 134 stores software and data, including, for example, but not limited to, testing software 136, datasets 138, predictive models 140, and model generator 142. As depicted in FIG. 1, in some embodiments, memory 134 of model generator 128 may include an artificial intelligence machine learning algorithm (AI 144).

[0027] 1 can be coupled to and control the operation of environmental event controller 126. Environmental event controller 126 of FIG. 1 can include several subsystems that generate controllable environmental events, such as, but not limited to, applying heat, applying cold, applying shock, applying vibration, etc.

[0028] As described above, in various embodiments, generator 128 operates DUT 102, for example via environmental event generator 126, to induce operational and environmental events that are likely to cause a change in frequency. When an event is applied, generator 128 causes the DUT to collect data 106 regarding changes in at least one frequency-related parameter. In some embodiments, collected data 106 is forwarded to generator 128 to create a data set of frequency-related parameters that correlate with operational and environmental events and parameters of those events (e.g., temperature, shock level, modem on time, positioning rate, etc.).

[0029] In this disclosure, the operation of test software 136 and DUT test software 124 as executed by processors 130 and 124, respectively, is described in more detail below with respect to Figure 2. The operation of model generator software 142 as executed by processor 130 is described in more detail below with respect to Figure 3.

[0030] In some embodiments of the present principles, the test signal to the modem 116 and / or GNSS receiver 112 may be simulated as a real-world signal such that the signal processing software 146 utilizes SUPERCORRELATION™ software 148 to monitor at least one frequency-related parameter in the same manner as the SUPERCORRELATION™ technique operates in a real-world environment, such as by processing signals received in a multipath environment.

[0031] In some embodiments, the predictive model generator 128 may be embedded in the DUT 102, with all signal processing performed by both the DUT 102 and the predictive model generator 128 being performed by the DUT 102. In such embodiments, the predictive model generation process may begin without a model, or alternatively, may begin with a model (e.g., a reference model) generated, for example, at the manufacturer or by a third party. The reference model may then be adapted and modified by the predictive model generator 128 as the DUT 102 is used under normal operating conditions. In such embodiments, because the DUT 102 is used in both good and poor signal reception environments, the SUPERCORRELATION™ technique may be used to determine at least one frequency-related parameter in accordance with present principles.

[0032] Alternatively, or in addition, in some embodiments, the predictive model generator 128 may be located remotely (i.e., in a cloud-based server) from the DUT 102 or DUTs that collect the frequency-related parameter data and environmental / operational event information. Processing of the frequency-related parameter data and environmental / operational event information in accordance with present principles may be performed to generate reference models for many different DUTs, or may be processed to generate a model for a specific DUT 102 that has been supplied with the specific frequency-related parameter data and environmental / operational event information.

[0033] 2 depicts a flow diagram of a method 200 for generating a dataset of test data in accordance with at least one embodiment of the present principles. Method 200 may be realized by execution of test software 136 and / or DUT test software 124 of FIG. 1. In embodiments where predictive model generator 128 is a standalone computing device coupled to DUT 102, test software 136 and DUT test software 124 may communicate with each other to generate the dataset. In some embodiments, where predictive model generator 128 is embedded in the DUT, the functionality of test software 136 and DUT test software 124 may be combined into a single process and executed on the DUT to generate the dataset.

[0034] 2, method 200 may begin at 202 and proceed to 204, during which test parameters are applied to the DUT. As described above, in some embodiments, the test parameters control various operations of the DUT, such as, but not limited to, display activity, display brightness, modem activity, CPU activity, crystal age, battery voltage, as well as combinations of operations. Method 200 may proceed to optional 206 or 208.

[0035] Similar to 204, at 206, an environmental event is applied to the DUT, such as, but not limited to, heating, cooling, shaking, dropping, etc. That is, in some embodiments, the predictive model generator of the present principles sends control parameters to an environmental event controller to generate an environmental event, including the severity and duration of the environmental event. Method 200 may proceed to 208.

[0036] At 208, frequency-related parameter data is collected as operations and / or combinations of operations (i.e., from 204 and 206 above) are performed on the DUT. As described above, in some embodiments, the frequency-related parameter data may include, for example, frequency and frequency rate information related to the operation of a frequency source of a UDT. In accordance with present principles, frequency is defined as the frequency source output oscillation rate (e.g., cycles per second), and frequency rate is defined as the rate of change of frequency over time. In some embodiments, other higher-order frequency-related parameters may also be used in accordance with present principles. Additionally, in some embodiments, the frequency-related parameter data may include the temperature of the frequency source during operation.

[0037] As mentioned above, in some embodiments, determining the at least one frequency-related parameter of interest can be performed using a SUPERCORRELATION™ technique to enable signal processing in poor signal reception environments, such as urban canyons where multipath interference is high. For example, the at least one frequency-related parameter can relate to signal processing of GNSS signals from multiple satellites. These signals are highly susceptible to multipath interference, but are important for generating an accurate frequency source in the DUT of the present principles. The SUPERCORRELATION™ technique uses a robust receiver motion compensation and co-correlation process to enable reception of GNSS signals in poor reception environments and to generate the accurate frequency-related parameter using those GNSS signals. Referring back to FIG. 2 , method 200 can proceed to 210.

[0038] At 210, method 200 constructs a dataset from the collected and determined data. That is, in some embodiments, the dataset includes a compilation of all collected data, including, but not limited to, actions, events, durations of actions and / or events, severity of actions and / or events, temperature information, frequency-related parameter information, time, etc. In some embodiments, the data may be collected at the DUT through execution of DUT test software. In such embodiments, the data may be collected in a file and transferred to, for example, a predictive model generator. Alternatively, or in addition, in some embodiments, the data may be streamed to the predictive model generator.

[0039] The method 200 may end at 212.

[0040] In some embodiments, method 200 may further include generating data when the DUT is used in the field. In such embodiments, the DUT may passively monitor operational activities and / or environmental events rather than actively controlling the occurrence of these events. In such embodiments, sensors such as accelerometers, gyroscopes, temperature sensors, etc. may be used to monitor the environment, and an operating system of the DUT may provide operational information about the DUT.

[0041] The collected data can then be used to train a model to predict the frequency response / change of the frequency source due to changes in at least one of the operating characteristics of the DUT or the environment in which the DUT is operating. For example, Figure 3 depicts a flow diagram of a method 300 of using a dataset to create a predictive model in accordance with at least one embodiment of the present invention. Method 300 can begin at 302 and proceed to 304, during which the dataset described above in Figure 2 is collected / accessed, for example, from memory accessible to an associated device. Method 300 can proceed to 306.

[0042] At 306, a determination is made as to what type of predictive model to create. For example, as depicted in method 300 of Figure 3, in some embodiments of the present principles, a user of an associated device may be provided with an opportunity to input a request, e.g., via the device's input / output devices, as to the type of model (e.g., LUT, AI / machine learning model, etc.) that the user would like generated from the collected / stored data. Method 300 may proceed to 308.

[0043] At 308, the collected / accessed data, including at least one of data relating to changes in at least one of the operating characteristics of the associated device or changes in characteristics of the environment in which the device is operating, and the respective frequency responses to the changes are correlated to generate a frequency response model. For example, in some embodiments, a look-up table (LUT) can be generated to correlate data parameters (e.g., event parameters) of at least one of the operating characteristics of the device and / or characteristics of the environment in which the device is operating with, for example, respective frequency-related parameters of the frequency source.

[0044] A generated model of the present principles, such as a LUT, can include data regarding environmental and / or operational events and associated parameters as inputs and can include frequency-related parameters as outputs. For example, a LUT input event can include activation of a display associated with an associated device, and an output can include a resulting frequency of a frequency source (e.g., oscillator) associated with the device, such as an increase in frequency. In this example, a known increase in frequency to the associated frequency source can be used by the DUT to modify signal processing to compensate for the change in frequency. In other embodiments, a LUT input can include an increase in environmental temperature causing a frequency increase. In other embodiments, an impact to the device may cause the frequency of the associated frequency source to "ring" with an oscillating change in frequency following the initial increase. All of these frequency-related parameter changes can be compensated for by the associated device, for example, in signal processing.

[0045] Returning to the method of FIG. 3 , method 300 may proceed to 310. At 310, the generated predictive model, such as a LUT, may be output and used, for example, to correct frequency-related parameter errors of a frequency source of an associated device. In some embodiments, the generated predictive model of the present principles may be used to anticipate behavior of a frequency-related parameter of a frequency source and compensate for that behavior during signal processing of received wireless signals (e.g., GNSS signals, WiFi signals, Bluetooth signals, etc.). The predictive model of the present principles finds particular use in commonly assigned U.S. patent application Ser. No. 63 / 424,185, filed Nov. 10, 2022, entitled “Method and Apparatus for Determining a Frequency Related Parameter of a Frequency Source Using Predictive Control.”

[0046] 3, in some embodiments of method 300, the collected / stored data can be implemented to train a machine learning neural network (artificial intelligence (AI)) to create a predictive model of the present principles. For example, if at 306 it is determined to generate a machine learning (artificial intelligence (AI)) model, method 300 can proceed to 312.

[0047] The machine learning neural network is executed at 312. The method 300 may proceed to 314.

[0048] At 314, the neural network is trained, for example, using the collected / accessed data set of FIG. 2. In some embodiments, the trained model has as input data associated with at least one of an operational event and / or an environmental event, as described above, and as output frequency-related parameters, e.g., a change to the frequency of a frequency source associated with an associated device. For example, in some embodiments, the input event may include activation of a display associated with the device, and the output may include the frequency of the frequency source (e.g., an oscillator) increasing in frequency. Information about the resulting frequency increase may then be used by the device associated with the frequency source to modify signal processing of the received signal to compensate for the change in frequency. Alternatively, or in addition, in some embodiments, the input event may include an increase in environmental temperature causing an increase in the frequency of a frequency source associated with a receiving device. In some embodiments, the input event may include an impact to the device, which may cause the frequency of the frequency source to "ring" in an oscillating change in frequency following the initial increase in frequency. In such and other similar embodiments, changes in such frequency-related parameters may be compensated for in signal processing, for example, by a receiving device associated with the frequency source.

[0049] That is, in some embodiments of the present principles, data corresponding to at least one operating characteristic of a wireless signal receiving device is collected while the device is in operation. A first training set is created including the collected data corresponding to the at least one operating characteristic of the wireless signal receiving device. In addition, data corresponding to at least one environmental event associated with the wireless signal receiving device is collected while the device is in operation. A second training set is created including the collected data corresponding to the at least one environmental event associated with the wireless signal receiving device. Furthermore, data corresponding to at least one frequency-related parameter corresponding to at least one of the at least one operating characteristic and / or the at least one environmental event associated with the wireless signal receiving device is collected, wherein the at least one frequency-related parameter changes with changes in the at least one operating characteristic and / or the at least one environmental event. A third training set is created including the collected data corresponding to the at least one frequency-related parameter. In such an embodiment, the neural network is trained in accordance with present principles using the first training set, the second training set, and the third training set to generate a predictive model used to compensate for changes in at least one operating characteristic and / or at least one frequency-related parameter corresponding to at least one environmental event.

[0050] Returning to method 300 of FIG. 3 , method 300 may proceed to 316. At 316, method 300 outputs a machine learning (e.g., AI-based) predictive model. The model can then be used to correct frequency-related parameter errors in a predictive manner based on current operational or environmental events affecting the frequency source function. That is, in some embodiments, a predictive model of the present principles can be output to be used, for example, to correct frequency-related parameter errors of a frequency source of an associated device. In some embodiments, the generated predictive model of the present principles can be used to anticipate behavior of frequency-related parameters of a frequency source and compensate for that behavior during signal processing of received wireless signals (e.g., GNSS signals, WiFi signals, Bluetooth signals, etc.). Alternatively, or in addition to modifying signal processing, the predictive model of the present principles can be used to compensate for frequency-related parameter errors by modifying the operation of the device, including but not limited to, turning the display off / on, modifying CPU activity, turning the modem on / off, etc., and / or by adjusting environmental characteristics of the environment in which the device associated with the frequency source is operating, including but not limited to, changing the temperature and / or humidity of the environment.

[0051] An embodiment of the predictive model of the present principles generated in accordance with 312-316 above finds particular use in U.S. patent application Ser. No. 63 / 424,185, filed Nov. 10, 2022, entitled "Method and Apparatus for Determining a Frequency Related Parameter of a Frequency Source Using Predictive Control."

[0052] The method 300 may end at 318.

[0053] In some embodiments, organizing a collected / stored dataset in accordance with present principles may involve processing the data, including, but not limited to, averaging, combining, filtering, and / or limiting the data within the dataset. Alternatively, or in addition, data within the dataset may be eliminated or corrected if the data is deemed irrelevant, redundant, or inaccurate.

[0054] In some embodiments of the present principles, a method for generating a predictive model includes operating a device under test while collecting data corresponding to at least one frequency-related parameter, at least one operating characteristic, and / or at least one environmental event, where the at least one frequency-related parameter changes with changes in the at least one operating characteristic and / or the at least one environmental event, and processing the data to generate a predictive model used to compensate for changes in the at least one frequency-related parameter corresponding to the at least one operating characteristic and / or the at least one environmental event.

[0055] In some embodiments of the method, processing the data may include correlating at least one of the at least one operating characteristic and / or the at least one environmental event with a respective at least one frequency-related parameter to determine a frequency response of a frequency source associated with the device under test to changes in at least one of the at least one operating characteristic and / or the at least one environmental event.

[0056] In some embodiments, the method may further include using the predictive model to correct a frequency response of a frequency source associated with the device under test for changes in at least one of the at least one operating characteristic and / or the at least one environmental event.

[0057] In some embodiments, in the method, correcting the frequency response includes at least one of adjusting for changes in frequency of the frequency source during signal processing of a signal received by the associated device under test, adjusting an operating characteristic of the device under test, and / or adjusting an environmental characteristic of an environment in which the device under test is operating.

[0058] In some embodiments, in the method, the predictive model comprises at least one of a lookup table and / or a machine learning model.

[0059] In some embodiments, in the method, the machine learning model includes a neural network that is trained using collected data corresponding to at least one frequency-related parameter, at least one operating characteristic, and / or at least one environmental event.

[0060] In some embodiments, in the method, the at least one operating characteristic includes at least one of a display activity of a display associated with the device under test, a display brightness of a display associated with the device under test, a modem activity of a modem associated with the device under test, a CPU activity of a CPU associated with the device under test, a crystal age of a crystal associated with the device under test, and / or a battery voltage of a battery associated with the device under test.

[0061] In some embodiments, in the method, the at least one environmental event includes at least one of heating, cooling, shaking, and / or dropping the device under test.

[0062] In some embodiments of the present principles, an apparatus for generating a predictive model includes at least one processor and at least one memory including at least one of programs and / or instructions that, when executed by the processor, cause the apparatus to perform operations including: operating a device under test while collecting data corresponding to at least one frequency-related parameter, at least one operating characteristic, and / or at least one environmental event, wherein the at least one frequency-related parameter changes with changes in the at least one operating characteristic and / or the at least one environmental event; and processing the data to generate a predictive model used to compensate for changes in the at least one frequency-related parameter corresponding to the at least one operating characteristic and / or the at least one environmental event.

[0063] In some embodiments, in the apparatus, processing the data includes correlating at least one of the at least one operating characteristic and / or the at least one environmental event with a respective at least one frequency-related parameter to determine a frequency response of a frequency source associated with the device under test to a change in at least one of the at least one operating characteristic and / or the at least one environmental event.

[0064] In the apparatus, the apparatus is further configured to compensate a frequency response of a frequency source associated with the device under test for changes in at least one of the at least one operating characteristic and / or the at least one environmental event using the predictive model.

[0065] In some embodiments, in the apparatus, correcting the frequency response includes at least one of adjusting for changes in the frequency of a frequency source during signal processing of a signal received by an associated device under test, adjusting an operating characteristic of the device under test, and / or adjusting an environmental characteristic of an environment in which the device under test is operating.

[0066] In some embodiments, in the device, the predictive model comprises at least one of a lookup table and / or a machine learning model.

[0067] In some embodiments, in the device, the machine learning model includes a neural network that is trained using collected data corresponding to at least one frequency-related parameter, at least one operating characteristic, and / or at least one environmental event.

[0068] In some embodiments, in the apparatus, the at least one operating characteristic includes at least one of a display activity of a display associated with the device under test, a display brightness of a display associated with the device under test, a modem activity of a modem associated with the device under test, a CPU activity of a CPU associated with the device under test, a crystal age of a crystal associated with the device under test, and / or a battery voltage of a battery associated with the device under test.

[0069] In some embodiments, in the apparatus, the at least one environmental event includes at least one of heating, cooling, shaking, and / or dropping the device under test.

[0070] In some embodiments, a computer-implemented method for training a neural network to generate a predictive model that compensates for frequency-related errors includes collecting data corresponding to at least one operating characteristic of a wireless signal receiving device in operation, creating a first training set including the collected data corresponding to the at least one operating characteristic of the wireless signal receiving device, collecting data corresponding to at least one environmental event associated with the wireless signal receiving device in operation, creating a second training set including the collected data corresponding to the at least one operating characteristic and / or the at least one environmental event associated with the wireless signal receiving device, collecting data corresponding to at least one frequency-related parameter corresponding to at least one of the first and second environmental events, wherein the at least one frequency-related parameter changes with changes in the at least one operating characteristic and / or the at least one environmental event; creating a third training set including the collected data corresponding to the at least one frequency-related parameter; and training a neural network using the first, second, and third training sets to generate a predictive model used to compensate for changes in the at least one frequency-related parameter corresponding to the at least one operating characteristic and / or the at least one environmental event.

[0071] In such embodiments, the generated predictive model may be used to correct a frequency response of a frequency source associated with the wireless signal receiving device for changes in at least one of the at least one operating characteristic and / or the at least one environmental event. In such embodiments, correcting the frequency response may include at least one of adjusting for changes in the frequency of the frequency source during signal processing of the signal received by the wireless signal receiving device, adjusting the operating characteristics of the wireless signal receiving device, and / or adjusting environmental characteristics of the environment in which the wireless signal receiving device is operating.

[0072] Those skilled in the art will understand that while various items are illustrated as being stored in memory or on a storage device during use, these items, or portions thereof, can be transferred between memory and other storage devices for purposes of memory management and data integrity. Alternatively, in other embodiments, some or all of the software components can execute in memory on another device and communicate with the illustrated computer system via computer-to-computer communications. Some or all of the system components or data structures can also be stored (e.g., as instructions or structured data) on a computer-accessible medium or on a portable article read by an appropriate drive, various examples of which are described above. In some embodiments, instructions stored on a separate computer-accessible medium can be transmitted to a computing device via a transmission medium or a signal, such as an electrical, electromagnetic, or digital signal, conveyed over a communications medium, such as a network and / or a wireless link. Various embodiments can further include receiving, sending, or storing instructions and / or data implemented in accordance with the foregoing description on a computer-accessible medium or via a communications medium. Generally, a computer-accessible medium may include a storage medium or memory medium, such as, for example, a magnetic or optical medium, such as a disk or DVD / CD-ROM, a volatile or non-volatile medium, such as RAM (e.g., SDRAM, DDR, RDRAM, SRAM, etc.), a ROM, etc.

[0073] The methods and processes described herein may be implemented in software, hardware, or a combination thereof in different embodiments. In addition, the order of methods may be changed, and various elements may be added, rearranged, combined, omitted, or otherwise modified. All examples described herein are presented in a non-limiting manner. Various modifications and variations may be made, as would be apparent to one of ordinary skill in the art having the benefit of this disclosure. Implementations according to embodiments have been described in the context of specific embodiments. These embodiments are meant to be illustrative, not limiting. Many variations, modifications, additions, and improvements are possible. Accordingly, components described herein as a single instance may be provided with multiple instances. Boundaries between various components, operations, and data stores are somewhat arbitrary, and particular operations are illustrated in the context of specific illustrative configurations. Other allocations of functionality are contemplated and may be within the scope of the following claims. Structures and functions presented as separate components in illustrative configurations may be implemented as combined structures or components. These and other variations, modifications, additions, and improvements may be within the scope of the embodiments, as defined by the following claims.

[0074] In the foregoing description, numerous specific details, examples, and scenarios are set forth to provide a more complete understanding of the present disclosure. However, it will be understood that embodiments of the present disclosure can be practiced without such specific details. Furthermore, such examples and scenarios are provided for illustrative purposes and are not intended to limit the present disclosure in any way. Together with the included description, one skilled in the art should be able to implement appropriate functionality without undue experimentation.

[0075] References herein to "an embodiment" or the like indicate that the described embodiment may include a particular feature, structure, or characteristic, but that not all embodiments necessarily include that particular feature, structure, or characteristic. Such phrases do not necessarily refer to the same embodiment. Furthermore, when a particular feature, structure, or characteristic is described in connection with an embodiment, it is believed to be within the knowledge of one skilled in the art to affect such feature, structure, or characteristic in connection with other embodiments, whether or not explicitly stated.

[0076] Embodiments according to the present disclosure may be implemented in hardware, firmware, software, or any combination thereof. Embodiments may also be implemented as instructions stored using one or more machine-readable media, which may be read and executed by one or more processors. A machine-readable medium may include any mechanism for storing or transmitting information in a form readable by a machine (e.g., a computing device or a "virtual machine" running on one or more computing devices). For example, a machine-readable medium may include any suitable form of volatile or non-volatile memory.

[0077] Additionally, the various operations, processes, and methods disclosed herein may be embodied in machine-readable media and / or machine-accessible media / storage devices compatible with a data processing system (e.g., a computer system) and may be performed in any order (e.g., using means to accomplish various operations). Accordingly, the specification and drawings should be regarded in an illustrative sense, and not in a restrictive sense. In some embodiments, the machine-readable media may be a non-transitory form of machine-readable media / storage device.

[0078] Modules, data structures, etc. defined herein are so defined for ease of discussion and are not intended to imply that any specific implementation details are required. For example, any of the described modules and / or data structures can be combined or divided into sub-modules, sub-processes, or other units of computer code or data as may be required by a particular design or implementation.

[0079] In the figures, a specific arrangement or order of schematic elements may be shown for ease of explanation. However, such a specific order or arrangement of elements does not imply that a particular order or sequence of operations or separation of processes is required in all embodiments. In general, schematic elements used to represent instruction blocks or modules may be implemented using any suitable form of machine-readable instructions, and each such instruction may be implemented using any suitable programming language, library, application programming interface (API), and / or other software development tool or framework. Similarly, schematic elements used to represent data or information may be implemented using any suitable electronic device or data structure. Furthermore, some connections, relationships, or associations between elements may be simplified or not shown in the figures so as not to obscure the present disclosure.

[0080] The present disclosure is to be considered illustrative and not restrictive in character, and all changes and modifications that come within the guidelines of the present disclosure are desired to be protected.

Claims

1. 1. A method for generating a predictive model, comprising: operating the device under test while collecting data corresponding to at least one frequency-related parameter, at least one operating characteristic, and / or at least one environmental event, wherein the at least one frequency-related parameter varies with changes in the at least one operating characteristic and / or the at least one environmental event; and processing the data to generate a predictive model used to compensate for the change in the at least one operating characteristic and / or the at least one frequency-related parameter corresponding to the at least one environmental event.

2. 2. The method of claim 1, wherein processing the data includes correlating at least one frequency-related parameter with at least one respective operating characteristic and / or at least one respective environmental event to determine a frequency response of a frequency source associated with the device under test to the change in at least one of the at least one operating characteristic and / or the at least one environmental event.

3. 3. The method of claim 1, further comprising using the predictive model to correct a frequency response of a frequency source associated with the device under test for the change in the at least one of the at least one operating characteristic and / or the at least one environmental event.

4. The method of claim 3 further comprising adjusting the frequency of the frequency source using a motion compensated correlation technique.

5. The method of claim 4 , wherein the motion compensated correlation technique comprises a hypercorrelation technique.

6. 4. The method of claim 3, wherein correcting the frequency response comprises at least one of adjusting for changes in frequency of the frequency source during signal processing of a signal received by the associated device under test, adjusting an operating characteristic of the device under test, and / or adjusting an environmental characteristic of an environment in which the device under test is operating.

7. The method of claim 1 , wherein the predictive model comprises at least one of a lookup table and / or a machine learning model.

8. 6. The method of claim 5, wherein the machine learning model comprises a neural network that is trained using the collected data corresponding to the at least one frequency-related parameter, the at least one operating characteristic, and / or the at least one environmental event.

9. 9. The method of claim 1, wherein the at least one operational characteristic comprises at least one of: display activity of a display associated with the device under test; display brightness of a display associated with the device under test; modem activity of a modem associated with the device under test; CPU activity of a CPU associated with the device under test; crystal age of a crystal associated with the device under test; and / or battery voltage of a battery associated with the device under test; and the at least one environmental event comprises at least one of heating, cooling, shaking, and / or dropping the device under test.

10. 1. An apparatus for generating a predictive model, comprising: at least one processor; and at least one memory containing at least one of a program and / or instruction, which, when executed by the processor, causes the device to: operating the device under test while collecting data corresponding to at least one frequency-related parameter, at least one operating characteristic, and / or at least one environmental event, wherein the at least one frequency-related parameter varies with changes in the at least one operating characteristic and / or the at least one environmental event; and processing the data to generate a predictive model used to compensate for the change in the at least one operating characteristic and / or the at least one frequency-related parameter corresponding to the at least one environmental event.

11. 11. The apparatus of claim 10, wherein processing the data includes correlating at least one frequency-related parameter with at least one respective operating characteristic and / or at least one respective environmental event to determine a frequency response of a frequency source associated with the device under test to the change in at least one of the at least one operating characteristic and / or the at least one environmental event.

12. The device comprises:

12. The apparatus of claim 10 or 11, further comprising: using the predictive model to correct a frequency response of a frequency source associated with the device under test for the change in the at least one of the at least one operating characteristic and / or the at least one environmental event.

13. The device comprises: The apparatus of claim 12 , further comprising: adjusting the frequency of the frequency source using a motion compensated correlation technique.

14. The apparatus of claim 13 , wherein the motion compensated correlation technique comprises a hypercorrelation technique.

15. 13. The apparatus of claim 12, wherein correcting the frequency response includes at least one of adjusting for changes in frequency of the frequency source during signal processing of a signal received by the associated device under test, adjusting an operating characteristic of the device under test, and / or adjusting an environmental characteristic of an environment in which the device under test is operating.

16. 16. The apparatus of claim 10, 11, 13, 14, and 15, wherein the predictive model comprises at least one of a lookup table and / or a machine learning model.

17. 17. The apparatus of claim 16, wherein the machine learning model comprises a neural network that is trained using the collected data corresponding to the at least one frequency-related parameter, the at least one operating characteristic, and / or the at least one environmental event.

18. 18. The apparatus of claim 10, wherein the at least one operational characteristic comprises at least one of: display activity of a display associated with the device under test; display brightness of a display associated with the device under test; modem activity of a modem associated with the device under test; CPU activity of a CPU associated with the device under test; crystal age of a crystal associated with the device under test; and / or battery voltage of a battery associated with the device under test; and the at least one environmental event comprises at least one of heating, cooling, shaking, and / or dropping the device under test.

19. 1. A computer-implemented method for training a neural network to generate a predictive model that compensates for frequency-related errors, comprising: collecting data corresponding to at least one operating characteristic of the wireless signal receiving device during operation; creating a first training set including the collected data corresponding to the at least one operating characteristic of the wireless signal receiving device; collecting data corresponding to at least one environmental event associated with the wireless signal receiving device during operation; creating a second training set including the collected data corresponding to the at least one environmental event associated with the wireless signal receiving device; collecting data corresponding to at least one frequency-related parameter corresponding to at least one of the at least one operating characteristic and / or the at least one environmental event associated with the wireless signal receiving device, the at least one frequency-related parameter changing with changes in the at least one operating characteristic and / or the at least one environmental event; creating a third training set including the collected data corresponding to the at least one frequency-related parameter; training the neural network using the first training set, the second training set, and the third training set to generate a predictive model used to compensate for the change in the at least one operating characteristic and / or the at least one frequency-related parameter corresponding to the at least one environmental event.

20. 20. The method of claim 19, further comprising using the predictive model to correct a frequency response of a frequency source associated with the wireless signal receiving device for the change in the at least one of the at least one operating characteristic and / or the at least one environmental event, wherein correcting the frequency response comprises at least one of adjusting for frequency changes of the frequency source during signal processing of signals received by the wireless signal receiving device, adjusting operating characteristics of the wireless signal receiving device, and / or adjusting environmental characteristics of an environment in which the wireless signal receiving device is operating.

Images