Adaptive Antenna Tuning Assisted by Machine Learning

Abstract

Certain aspects of the present disclosure provide techniques for adaptively tuning a wireless data transmission system in an electronic device that includes receiving one or more operating characteristics of a wireless data transmission system of the device, determining a target wireless data transmission system configuration based on the one or more operating characteristics using a wireless data transmission system configuration model, and implementing the target wireless data transmission system configuration in the wireless data transmission system.

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to U.S. Application No. 17 / 463,290, filed August 31, 2021, which is assigned to the assignee of this application and is incorporated by reference in its entirety into this specification.

[0002] introduction Aspects of the present disclosure relate to adaptive antenna tuning.

[0003]

[0003] A wireless device such as a smartphone may have one or more antennas for transmitting and receiving wireless data by a wireless data transmission system. How such a device is used, for example, how the smartphone is held, may affect the performance of the antenna and therefore the performance of the wireless data transmission. Degraded antenna performance may lead to slower wireless data transmission, increased battery usage, and increased wireless network interference, to name a few problems.

[0004]

[0004] Therefore, what is needed is a system and method for dynamically adapting antenna tuning to improve wireless device performance. Summary of the Invention

[0005]

[0005] Certain aspects provide a method for adaptively tuning a wireless data transmission system in an electronic device, the method including receiving one or more operating characteristics of the wireless data transmission system of the device, determining a target wireless data transmission system configuration based on the one or more operating characteristics using a wireless data transmission system configuration model, and implementing the target wireless data transmission system configuration in the wireless data transmission system.

[0006]

[0006] Another aspect provides a processing system configured to perform the aforementioned methods and methods further described herein; a non-transitory computer readable medium comprising instructions which, when executed by one or more processors of the processing system, cause the processing system to perform the aforementioned methods and methods further described herein; a computer program product embodied on the computer readable storage medium comprising code for performing the aforementioned methods and methods further described herein; and a processing system comprising means for performing the aforementioned methods and methods further described herein.

[0007] The following description and the related drawings set forth in detail certain illustrative features of the one or more aspects. [Brief description of the drawings]

[0008]

[0008] The accompanying drawings illustrate some aspects of the one or more aspects and therefore should not be considered as limiting the scope of the present disclosure. [Figure 1A]

[0009] FIG. 1 illustrates example test results for three different use cases of a wireless electronic device. [Figure 1B]

[0010] FIG. 1 illustrates an example use case decision boundary for a wireless electronic device based on measured test results. [Diagram 2]

[0011] FIG. 1 illustrates an example system for performing adaptive antenna tuning in a wireless electronic device. [Diagram 3]

[0012] FIG. 1 illustrates an exemplary reinforcement learning model architecture. [Figure 4]

[0013] FIG. 2 illustrates an exemplary performance estimator. [Diagram 5]

[0014] 1 illustrates an exemplary method for performing adaptive antenna tuning. [Figure 6]

[0015] FIG. 1 illustrates an example electronic device that may be configured to perform adaptive antenna tuning as described herein.

[0009]

[0016] For ease of understanding, wherever possible, like reference numbers have been used to designate like elements common to the figures. It is contemplated that elements and features of one embodiment may be beneficially incorporated in other embodiments without further recitation. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010]

[0017] Aspects of the present disclosure provide apparatus, methods, processing systems, and computer-readable media for machine learning assisted adaptive antenna tuning.

[0011]

[0018] As wireless electronic devices become smaller and more capable, the problem of designing wireless data transmission systems becomes more onerous, due at least in part to the complexity of designing effective antennas within the tight and complex interior confines of such devices. Indeed, modern wireless electronic devices can operate simultaneously at many wireless frequencies for both transmission and reception, often requiring multiple antennas of different designs within a single device.

[0012]

[0019] Functional design and packaging issues are only one aspect of the overall problem of creating an effective wireless data transmission system for a wireless electronic device. Different modes of use of such a device pose other problems, as different modes affect the performance of the wireless data transmission system in different ways.

[0013]

[0020] For example, holding a wireless electronic device such as a smartphone or tablet computer in the left or right hand, or in both hands, may change wireless data transmission performance because different hand placements affect different antennas differently. As another example, placing it on a surface, placing it in an area (e.g., in a pocket, car, or airplane), using it plugged in (e.g., via a USB or similar cable), etc., generally affect wireless data transmission performance differently.

[0014]

[0021] Different approaches can be used to improve antenna performance in wireless communication enabled devices, for example, impedance matching and tuning can be performed dynamically to improve the performance of antennas in such devices.

[0015]

[0022] In general, impedance opposes the flow of energy through a system. A constant electronic signal may have a constant impedance, and a varying electronic signal may have an impedance that varies with changing frequency. Impedance generally has a complex value with a resistive component forming the "real" part of the value and a reactive component forming the "imaginary" part of the value.

[0016]

[0023] Antenna impedance relates to the voltage and current at the input to the antenna. The "real" part of the antenna impedance represents the power dissipated or absorbed from the antenna, and the "imaginary" part of the impedance represents the power stored in the near field of the antenna, i.e., the non-radiated power. Antennas are generally more efficient, and therefore more effective, when the impedance of the system is optimized for the antenna.

[0017]

[0024] Impedance matching refers to designing the input impedance of an electrical load or the corresponding output impedance of a signal source to maximize power transfer and minimize signal reflection from the load. However, since impedance varies with changes in frequency signal, dynamic impedance tuning may be used to tune an antenna to match a radio frequency (RF) front end so that power transfer from the RF front end to the antenna is maximized.

[0018]

[0025] Impedance tuning can be performed in an "open loop" configuration, where preconfigured parameters are used to tune the antenna to a system, or in a "closed loop" configuration, where parameters are dynamically adjusted to tune the antenna to a system. In either case, such parameters can be used to improve antenna performance (e.g., total radiated power and return loss) of a wireless electronic device.

[0019]

[0026] Aperture tuning is another method to improve the performance of wireless communication systems in electronic devices. In general, antenna aperture tuning involves modifying the resonant frequency of an antenna to match a specific application or frequency. By tuning the resonant frequency of an antenna for a specific application or frequency, the efficiency of the antenna is improved for that specific application or at that specific frequency. Aperture tuning can therefore allow the same antenna to be used more efficiently for multiple applications at multiple frequencies.

[0020]

[0027] One way to perform antenna aperture tuning is to modify the electrical length of the antenna to adjust its resonant frequency. In some embodiments, a switch can be used to adjust the resonant frequency of the antenna by connecting the antenna to a ground path of different length, thus shifting the antenna resonance and changing the antenna performance without any structural changes to the antenna. A capacitor or inductor can also be used to further adjust the resonant frequency and can generally be connected between the switch and the radiating element of the antenna.

[0021]

[0028] Aperture tuning and impedance tuning can generally improve the performance of an antenna, such as its operating band, return loss, bandwidth, gain, and efficiency. In modern mobile devices, such as smartphones, tablet computers, and smart wearables, aperture tuning can beneficially improve the device's ability to operate in multiple bands at different times, which may be referred to as band-selective tuning.

[0022]

[0029] Aspects described herein relate to systems and methods for performing adaptive antenna tuning in wireless communication systems. Unlike some conventional methods that rely on fixed antenna tuning configurations (e.g., based on a well-defined use case), the methods described herein relate to using reinforcement learning to teach an agent to select one or more optimal antenna tuner settings based on feedback data. The antenna tuner settings may generally relate to any type of tuner in a wireless device, including impedance tuners and / or aperture tuners in various examples described herein. Beneficially, the feedback data can be generated on-device by using an estimation model that compares a selected antenna tuner setting to a non-selected antenna tuner setting to determine the performance of the selection. In some aspects, further feedback data can be generated based on sensors in the device and feedback from a network that wirelessly communicates with the device.

[0023]

[0030] Thus, aspects described herein advantageously allow the device to learn its own optimal wireless communications device configurations, such as impedance tuner settings, aperture tuner settings, and other tuner settings, reducing the device's sensitivity to manufacturing changes and other external factors, such as how the device is used (e.g., how it is held) and how the device may be modified by a user (e.g., by placing a protective cover on the device). Furthermore, laboratory characterization of well-defined specific use cases is no longer required to achieve improved antenna performance.

[0024]

[0031] Beneficially, the aspects described herein generally improve antenna performance (e.g., by 3-5 dB or more) of devices having wireless communication systems, increase the transmission and reception range of such devices, increase battery life of battery-powered devices, and reduce interference in wireless data communication networks used by such devices.

[0025] Characterizing Conventional Use Cases for Antenna Aperture Tuning

[0032] Conventional antenna tuning is based on determining a particular use case configuration, such as left-handed or right-handed holding of the device.

[0026]

[0033] 1A shows example impedance values ​​for three different use cases (free space, right hand, and left hand) and the averages calculated for each set of test results. For example, test results 102A-C for a particular use case (in this example, an example right hand grasp) result in an averaged point 102D. The averaged point can be considered a characteristic or representative point for the use case based on the test results. In this example, the test results measure real and imaginary impedance for a given frequency, and the tests can be performed across many different frequencies based on the capabilities of the device being tested.

[0027]

[0034] Based on the measurement results, clear use case decision boundaries can be determined (as indicated by the arcs) as shown in Figure 1B. The "live" measurements 104 can then be plotted and compared to representative points for different use cases and / or to the use case decision boundaries to determine how the wireless electronic device is currently being used (e.g., the device's current use case).

[0028]

[0035] Unfortunately, initial modeling between measurements and use cases, such as that shown in Figure 1A, is time consuming, costly, and generally unrepresentative because it is impractical to test the myriad of different ways that any given device going into production may be used (e.g., for various use cases). For example, a device may be tested in several different positions to determine several different use cases, but the tests rarely truly represent the vast variety of different ways that the device may be handled by various users. For example, as shown in Figure 1B, the decision boundaries cover a large area around the representative points for each of the three example use cases, and those large areas may over-represent certain use cases.

[0029]

[0036] 1B, there are many areas where use case decision boundaries overlap, such that a point may be within the limits of multiple use cases, leading to further uncertainty in use case decisions. Thus, using a clean decision boundary based on representative points, as in this example, may require additional processing by the wireless electronic device, which slows down operation and uses more power.

[0030]

[0037] More generally, the limited nature of traditional testing leads to limited use-case test data and models that have limited ability to reliably determine how a device is currently being used (e.g., with respect to a defined set of use cases) based on live data measurements. In particular, misdetermination or misprediction of the use case may cause incorrect antenna tuner settings to be applied, leading to suboptimal wireless data transmission performance of the device.

[0031]

[0038] To improve upon conventional testing and use case determination, aspects described herein do not rely on use case determination as a prerequisite for determining optimal antenna tuner settings. Rather, aspects described herein dynamically determine optimal antenna tuner settings (e.g., of impedance tuners and / or aperture tuners) based on feedback data. In some aspects, an agent is taught by reinforcement learning an optimal policy for selecting antenna tuner settings based on various inputs, feedback signals, and a selection performance function configured to reinforce the selection of optimal antenna tuner settings. Beneficially, there is no need to determine the use case; rather, the device can adapt to any manner of use case based on a reinforcement learning feedback loop.

[0032] Introduction to Reinforcement / Agent Learning

[0039] Reinforcement learning is a form of machine learning that focuses on teaching an "agent" (also called a "smart agent" or "AI agent") what actions to take in an environment to maximize rewards and / or minimize penalties for the actions taken. Reinforcement learning generally differs from supervised learning in that it does not require labeled input / output pairs to train the agent, which is very useful in scenarios where it is difficult to collect sufficient training data for traditional supervised learning, such as those described above with respect to Figures 1A and 1B with respect to characterizing the use case. Reinforcement learning thus focuses on learning the best policy for an agent that maximizes a performance function, which may include, for example, a reward component that is maximized over time and / or a penalty component that is minimized over time.

[0033]

[0040] A reinforcement learning agent can interact with an environment at discrete time steps or time intervals. In general, at each time step, the agent receives a current state and evaluates a performance function that determines a reward and / or penalty based on the state. The agent then selects an action from a set of available actions, which is then sent to the environment. The environment transitions to a new state based on the action, and a new reward and / or penalty associated with the transition is determined and fed back to the agent. Thus, the ultimate goal of a reinforcement learning agent is to learn a policy that maximizes expected cumulative reward and / or minimizes expected cumulative penalty based on the selection of different actions at different times in a changing environment.

[0034]

[0041] The "policy" adopted by the agent can be based on a machine learning model, such as a neural network model. For example, various inputs can be provided to the machine learning model, and the model can predict an action to take based on the inputs. Reinforcement learning can vary the parameters (e.g., weights and biases) of such a model over time to improve the model and the agent's policy-based behavior based on a performance function.

[0035]

[0042] For example, in the context of antenna tuning, an agent may select (e.g., "behavior") an antenna tuner setting, such as an impedance tuner setting, an aperture tuner setting, and / or another tuner setting based on a policy (e.g., model) and provide it to a device (e.g., "environment") that is implemented for data transmission (e.g., change the state of the device). Transmission according to the selected antenna tuner setting may be used to generate feedback data, which may be used to determine a reward and / or penalty for the agent based on the antenna tuner setting selected for data transmission. In some aspects, a policy used by an agent may take into account various types of input data, such as a measured real impedance along a transmit chain of a wireless data communication system, an imaginary impedance along the transmit chain, one or more operating frequencies of the transmit chain, and one or more current antenna tuner settings (e.g., current state), such as a current impedance tuner state, a current aperture tuner state, and / or other current tuner states.

[0036] Exemplary System for Performing Adaptive Antenna Tuning - Patent application

[0043] FIG. 2 illustrates an example system 200 for performing adaptive antenna tuning in a wireless electronic device.

[0037]

[0044] System 200 includes a modem 210 , which in this example includes a measurement component 212 , an adaptive antenna tuning component 214 , a performance estimation component 216 , and an antenna tuner settings database 218 .

[0038]

[0045] The measurement component 212 may be configured to receive measurement data, such as impedance and frequency measurements, from other aspects of the wireless electronic device. For example, the measurement component 212, in one aspect, may receive measurements from a feedback receiver (FBRx) 222 of the wireless transceiver 220. In general, the feedback receiver 222 is a circuit that compares measurements of a transmitted signal at different points along the transmit chain. For example, a voltage standing wave ratio (VSWR) or return loss may be determined, which provides a measurement of the complex impedance of the transmitted signal. An aspect of the modem 210, such as the adaptive antenna tuning component 214, may then receive the complex impedance and convert it to an impedance at the antenna. As described above, this antenna impedance may be used to determine how the antenna is affected by different antenna tuner settings, such as impedance tuner settings, aperture tuner settings, and / or other tuner settings implemented by the impedance tuner 232, aperture tuner 234, and / or other tuner 236, respectively.

[0039]

[0046] The adaptive antenna tuning component 214 may implement an agent for determining antenna tuner settings, including settings of the impedance tuner 232 and / or the antenna aperture tuner 234, as further described with respect to FIG. 3. In some cases, the agent may determine and / or select from settings stored in the antenna tuner settings database 218. The settings may include aperture, impedance, and / or other tuner settings that may be implemented by the impedance tuner 232, the aperture tuner 234, and / or other tuners 236, which may be considered aspects of the state of the radio frequency front end (RFFE) 230 when implemented. For example, the state of the radio frequency front end 230 may be defined by settings of the impedance tuner 232, the aperture tuner 234, and / or other tuners 236.

[0040]

[0047] In this example, the radio frequency front end 230 includes an impedance tuner 232 and an aperture tuner 234, which are configured to perform impedance tuning and aperture tuning, respectively. The radio frequency front end 230 further includes another tuner 236, which may include, for example, one or more matching networks for matching the output of the power amplifier to an antenna to improve power amplifier performance.

[0041]

[0048] It should be noted that the radio frequency front end 230 may include many other aspects not shown in this example for simplicity, such as power amplifiers, power trackers, duplexers, hexaplexers, switches, low noise amplifiers, filters, antenna switches, and extractors, to name a few.

[0042]

[0049] The radio frequency front end 230 is further connected to an antenna 250. Note that in this example, a single antenna 250 is shown for simplicity, but the radio frequency front end 230 may be connected to multiple antennas. Additionally, while this example shows a single radio frequency front end, other examples may include multiple radio frequency front ends for different radio access technologies, for operating simultaneously at different frequencies, etc.

[0043]

[0050] The adaptive antenna tuning component 214 is further configured to provide antenna tuner settings to the impedance tuner 232, the aperture tuner 234 and / or other tuners 236 in the radio frequency front end 230 to improve the performance of the antenna 250 (which may represent multiple antennas). An example of an adaptive antenna tuning system that may be implemented by the adaptive antenna tuning component 214 is shown and described in more detail with respect to FIG.

[0044]

[0051] The adaptive antenna tuning component 214 can significantly improve the performance of the wireless data transmission system of the electronic device. For example, a 3-5 dB improvement in antenna efficiency can be achieved along with reduced power usage, increased battery life, and lower network interference. Such improvements are generally beneficial and may be particularly suitable for some scenarios, such as use indoors and / or at the cell edge, or when the device is served by multiple antennas for coverage.

[0045]

[0052] To transmit and receive data, the modem 210 is connected to a wireless transceiver 220, which is connected to a radio frequency front end 230, which is connected to an antenna 250. It should be noted that the modem 210 may include many other aspects not shown in this example for simplicity, such as a processing core, read only memory (ROM), random access memory (RAM), security components, peripheral components, cache, and more.

[0046]

[0053] In particular, for simplicity, FIG. 2 shows only selected aspects of the device's wireless data transmission system; many other aspects are possible, such as other processors, memories, sensors, input and output devices, peripheral systems, etc.

[0047] Example Reinforcement Learning Model Architecture

[0054] 3 illustrates an example reinforcement learning model architecture 300. The reinforcement learning model architecture 300 can be implemented, for example, by the adaptive antenna tuning component 214 of FIG.

[0048]

[0055] The reinforcement learning model architecture 300 includes an agent 304 that may implement policies for determining or selecting antenna tuner settings, which in this example include an impedance tuner setting for the impedance tuner 232 and an aperture tuner setting for the aperture tuner 234. Although not shown in FIG 3, it should be noted that in additional aspects, the agent 304 may also be configured to determine tuner settings for other tuners, such as the other tuner 236 described above with respect to FIG 2.

[0049]

[0056] In some cases, the agent 304 may implement its policies through a machine learning model (e.g., a “policy model” or “wireless data transmission system configuration model”) such as a neural network model that receives one or more inputs 302 (e.g., operational characteristics of a wireless communication system in the device) and outputs a policy decision (e.g., a target wireless data transmission system configuration) such as an impedance tuner setting 303 (“IT setting” in FIG. 3 ) and / or an aperture tuner setting 305 (“AT setting” in FIG. 3 ). In this example, one or more inputs 302 provided to the agent 304 may be used as inputs to the policy model of the agent 304 and include a real impedance measurement 302A, an imaginary impedance measurement 302B, one or more operating frequencies 302C, an aperture tuner state 302D, an impedance tuner state 302E, and modem metrics 302F (e.g., received signal received power (RSRP) metric, received signal received quality (RSRQ) metric, signal to noise metric (SNR / SINR), received signal strength indicator (RSSI), etc.). Additionally, although not shown, other input data may be used, such as sensor data from devices on which the wireless data communications system is operating.

[0050]

[0057] In some cases, the policy model implemented by the agent 304 may output one or more indexes related to antenna tuner settings, including impedance tuner settings, aperture tuner settings, and / or other tuners that may include any number of tuner parameters. In some cases, the indexes may identify tuner settings stored in a database, such as the settings database 218 of FIG.

[0051]

[0058] In some aspects, the agent 304 can be configured to further output the expected performance (efficiency, matching impedance) of different antenna tuner settings for one or more carriers present on the antenna and let other parts of the system (e.g., external to the adaptive antenna tuning system) select the tuner behavior based on other goals and constraints, such as power amplifier efficiency, nonlinearity, accuracy (e.g., as described by the magnitude of the error vector), achievable modulation type, achievable antenna rank for multiple-input multiple-output (MIMO) configurations, etc.

[0052]

[0059] As shown, complex impedance measurements (including real and imaginary components 302A and 302B, respectively) may be provided by impedance measuring component 308. In some aspects, a wireless data communication system may use multiple operating frequencies simultaneously, in which case impedance measurements by impedance measuring component 308 may be based on multiple simultaneous frequencies. However, it should be noted that in such cases, the measurement frequencies need not necessarily match all of the operating frequencies.

[0053]

[0060] In some cases, the impedance measurements may be generated by a feedback receiver, such as feedback receiver 222 of Figure 2. Additionally, the impedance tuner state and aperture tuner state may be known by agent 304 based on the current state of the wireless data transmission system or may otherwise be provided by impedance tuner 232 and aperture tuner 234 as a primary or validation reference.

[0054]

[0061] The antenna tuner settings (e.g., impedance tuner settings 303 and / or aperture tuner settings 305 in this example) determined by the agent 304 may be further provided to a performance estimator 306 configured to provide feedback data to the agent 304 based on the determined antenna tuner settings. In some cases, the aperture tuner settings 305 and the impedance tuner settings 303 can be collectively considered as a tuning code, and the agent 304 attempts to determine an optimal tuning code including optimal impedance tuner and aperture tuner settings that optimize the performance (efficiency, power, linearity) of the radio frequency front end (e.g., 230 in FIG. 2). In one example, maximum power delivery from the power amplifier to the antenna (e.g., 250 in FIG. 2) is achieved based on the optimal tuning code. In various aspects, maximum received power, minimum nonlinearity, and a tradeoff between transmit power, received power, and linearity are all considered in determining the optimal tuning code.

[0055]

[0062] In some aspects, the feedback data is based on an estimated impedance associated with an antenna tuner setting not selected by the agent 304 (e.g., in this example, an unselected or "virtual" impedance tuner setting and / or aperture tuner setting). For example, any number of defined impedance tuner settings and / or aperture tuner settings may be stored (e.g., in the settings database 218 of FIG. 2) and made accessible to the agent 304. The estimated impedance based on the antenna tuner setting output by the agent 304 may be based on one or more models implemented by the performance estimator 306, as further described with respect to the example of FIG. 4.

[0056]

[0063] In some aspects, the feedback data may additionally or alternatively be based on other indicators. As noted above, various sensors on the device may provide sensor data that may be used to generate feedback to the agent 304. Similarly, indicators (e.g., signal quality metrics) from a wireless data network in data communication with the device may be used to generate feedback data to the agent 304.

[0057]

[0064] Based on feedback data from the performance estimator 306, the agent 304 can update its policy (e.g., parameters of its policy model) using reinforcement learning. For example, the agent 304 can evaluate a selection performance function that rewards the agent 304 for selecting the best adaptive antenna tuning settings based on performance estimates from the performance estimator 306, which in this example relate to the impedance tuner 232 and the aperture tuner 234, but may also include other antenna tuners (e.g., other tuners 236 of FIG. 2) as described above. As a further example, the selection performance function can reward the selection of adaptive antenna tuner settings that minimize power loss at the antenna.

[0058]

[0065] The selection performance function evaluated by the agent 304 may additionally or alternatively penalize the agent 304 for selecting suboptimal adaptive antenna tuning settings. For example, the selection performance function may include a penalty factor for switching between tuner settings to discourage rapid cycling between tuner settings. Generally speaking, the selection performance function may include reward and / or penalty factors that cause the agent 304 to adapt its setting selection policy over time to maximize the reward factor while simultaneously minimizing the penalty factor.

[0059]

[0066] It should be noted that a device implementing the reinforcement learning model architecture 300 may enable or disable the feedback loop provided by the performance estimator 306 and reinforcement learning as needed. For example, a new device may be pre-configured with a policy model for selecting impedance and / or aperture tuner settings. Once in use, the new device may enable the feedback mechanism provided by the performance estimator 306 and reinforcement learning so that the policy model of the agent 304 improves over time. After some time, for example after the policy model has converged (or is presumed to have converged), the feedback loop may be disabled. Additionally, the feedback loop may be enabled at intervals to ensure that any changes to the device (e.g., a new phone cover is installed) are understood by the agent 304. When the feedback loop is disabled, the agent 304 continues to determine antenna tuner settings, such as the impedance tuner setting 303 and the aperture tuner setting 305, based on its existing policy without the feedback provided by the performance estimator 306.

[0060]

[0067] Beneficially, the reinforcement learning model architecture 300 can determine the aperture tuner settings (e.g., 305) without having to first determine the use case of the device, or at all. Thus, a device implementing the reinforcement learning model architecture 300 may be able to adaptively tune its aperture tuner settings based on any use case, without sensitivity to manufacturing differences between devices.

[0061] Exemplary Performance Estimator

[0068] Generally, a device having a wireless communication system can "see" only the impedance at its antenna, the frequency it is operating on for transmission, and its current antenna tuner state (e.g., impedance and / or aperture tuner state). In order to provide a feedback signal to an agent (e.g., agent 304 of FIG. 3), the device can estimate how the device would have performed using different antenna tuner settings.

[0062]

[0069] In some aspects, as described below, a machine learning based estimator can be used to simulate the performance of a wireless communication system with non-selected antenna tuner settings without physically switching to them, saving time and power. Additionally, an additional estimator, which may also be implemented by a machine learning model, can be used to estimate the efficiency of the antenna based on both the actual antenna tuner states and the measured actual performance, as well as the non-selected antenna tuner states and the estimated performance in those states. Once all of these values ​​are determined, the system calculates the optimal antenna tuner setting (which may be one of the non-selected antenna tuner settings) and provides this decision as feedback to the agent for learning.

[0063]

[0070] FIG. 4 illustrates an example performance estimator 306 that may be used to provide feedback data to an agent, such as agent 304 of FIG.

[0064]

[0071] In the example performance estimator 306 shown in FIG. 4, the antenna impedance estimator 402 estimates antenna tuner settings that are not currently being used by the device (e.g., AT1 to AT2 in this example). N-1 , where N is the total number of antenna tuner settings), predicts "virtual" impedance estimates (e.g., estimated operating characteristics of a wireless data transmission system) 404A-404B including real and imaginary components. In some aspects, the antenna tuner settings include impedance tuner and aperture tuner settings, while in other examples, the antenna tuner settings may relate to impedance tuners or aperture tuners. In this example, the current antenna tuner settings AT N may include a current aperture tuner state 302D and a current impedance tuner state 302E.

[0065]

[0072] In some aspects, the antenna impedance estimator 402 may be implemented using a machine learning model, such as a trained neural network model, a decision tree, a boosted tree, or another type of model, trained to predict the virtual impedance based on the aperture tuner setting.

[0066]

[0073] Based on the antenna impedance estimates 404A-404B, an antenna efficiency estimator 406 predicts virtual antenna efficiency estimates (e.g., estimated performance metrics) 408A-408B. The virtual antenna efficiency may be, for example, a function of different antenna tuner settings (e.g., AT1-AT2 in this example). N-1) In addition, the antenna efficiency estimator 406 determines the efficiency of the antenna based on the actual antenna tuner settings, as in the current aperture tuner state 302D and impedance tuner state 302E.

[0067]

[0074] Both the virtual antenna efficiency based on the non-selected antenna tuner settings (e.g., 408A-408B) and the actual antenna efficiency based on the current antenna tuner setting (e.g., 408C) are provided to a comparator 410 to determine an optimal antenna tuner setting 412. For example, the optimal aperture tuner setting 412 can be the setting that provides the highest transmit power efficiency. The optimal antenna tuner setting 412 can be provided to an agent, such as agent 304 of FIG. 3, as feedback data for performing reinforcement learning, as described above. Thus, the feedback data can be used to help the agent adapt its policy to select the optimal antenna tuner setting more frequently, for example, based on device operating characteristics.

[0068]

[0075] 4, other inputs may be considered by the performance estimator 306, and additional models may be used to estimate virtual performance values ​​that can be compared (e.g., by the comparator 410) to actual performance values ​​to determine the best antenna tuner setting. For example, the performance estimator 306 may consider indicators provided by a network in wireless data communication with the device, including a received signal received power (RSRP) metric, a received signal received quality (RSRQ) metric, a signal to noise metric (SNR / SINR), a received signal strength indicator (RSSI), to name a few. The actual values ​​provided by the network may be compared to estimates from a suitable model to select an optimal aperture tuner setting 412.

[0069] Exemplary Methods for Adaptive Antenna Tuning

[0076] FIG. 5 illustrates an example method 500 for performing adaptive antenna tuning in a device having a wireless data communication system, such as described above with respect to FIGS.

[0070]

[0077] Method 500 begins with receiving one or more operational characteristics of a wireless data transmission system of a device, at step 502. For example, the operational characteristics can be similar to the inputs 302 described with respect to FIG. 3 and others, such as sensor data from the device.

[0071]

[0078] The method 500 then proceeds to step 504, where the wireless data transmission system configuration model is used to determine a target wireless data transmission system configuration based on the one or more operating characteristics. For example, the model may be implemented by an agent, such as agent 304 described above with respect to FIG.

[0072]

[0079] The method 500 then proceeds to step 506, where the target wireless data transmission system configuration is implemented in the wireless data transmission system.

[0073]

[0080] In some aspects, the method 500 further includes determining one or more estimated operating characteristics of the wireless data transmission system using one or more virtual wireless data transmission system configurations that differ from the target wireless data transmission system configuration using a first estimator model, as described with respect to FIG 4 in conjunction with the impedance estimator 402. In some aspects, the first estimator model comprises a neural network model.

[0074]

[0081] In some aspects, the method 500 further includes determining one or more estimated performance metrics of the wireless data transmission system based on the one or more estimated operating characteristics using a second estimator model, as described with respect to FIG. 4 in conjunction with the antenna efficiency estimator 406. In some aspects, the second estimator model comprises a neural network model.

[0075]

[0082] In some aspects, the method 500 further includes determining an actual performance metric based on the implemented target wireless data transmission system configuration, as described with respect to FIG. 4 in conjunction with the antenna efficiency estimation 408C.

[0076]

[0083] In some aspects, the method 500 further includes determining a best wireless data transmission system configuration among the target wireless data transmission system configuration and the one or more virtual wireless data transmission system configurations based on the actual performance metric and the one or more estimated performance metrics, as described with respect to FIG. 4 in conjunction with the comparator 410.

[0077]

[0084] In some aspects, the method 500 further includes generating feedback data based on the determined best wireless data transmission system configuration, hi some aspects, the feedback data includes the best wireless data transmission system configuration as described with respect to FIG.

[0078]

[0085] In some aspects, the feedback data includes one of a reward signal for a reinforcement learning model if the target wireless data transmission system configuration matches the best wireless data transmission system configuration or a penalty signal if the target wireless data transmission system configuration does not match the best wireless data transmission system configuration.

[0079]

[0086] In some aspects, the method 500 further includes updating the wireless data transmission system configuration model based on the feedback data. hi some aspects, the wireless data transmission system configuration model is a reinforcement learning model.

[0080]

[0087] In some aspects, the actual performance metric includes an indicator provided by a network in wireless data communication with the device. In some aspects, the indicator includes at least one of a received signal received power (RSRP) metric, a received signal received quality (RSRQ) metric, a signal to noise metric (SNR / SINR), or a received signal strength indicator (RSSI). In other aspects, other network key performance indicators (KPIs), or even metrics crowdsourced from other devices on the network, can be used as the actual performance metric.

[0081]

[0088] In some aspects, the method 500 further includes determining an optimal tuning code for the impedance tuner based on the target wireless data transmission system configuration. Generally, the optimal tuning code is one for which there is a minimum loss between the power delivered from the power amplifier and the power transmitted by the antenna.

[0082]

[0089] In some aspects, the one or more operating characteristics include a real impedance of an element of the wireless data transmission system, an imaginary impedance of an element of the wireless data transmission system, one or more frequencies of the wireless data transmission system, and an antenna tuner state. For example, the antenna tuner state can include one or more of an impedance tuner state, an aperture tuner state, and / or other tuner states.

[0083]

[0090] In some aspects, the element is an antenna of a wireless data transmission system and the target wireless data transmission system configuration includes an antenna tuner setting, hi some aspects, the antenna tuner setting includes at least one of an aperture tuner setting or an impedance tuner setting.

[0084] Exemplary Electronic Device for Performing Adaptive Antenna Tuning - Patent application

[0091] FIG. 6 illustrates an example processing system 600 for performing sparsity-aware compute-in-memory, for example, as described herein with respect to FIGS.

[0085]

[0092] The processing system 600 includes a central processing unit (CPU) 602, which in some examples may be a multi-core CPU. Instructions executed in the CPU 602 may be loaded from a program memory associated with the CPU 602 or may be loaded from a memory partition 624, for example.

[0086]

[0093] The processing system 600 also includes additional processing components tailored to specific functions, such as a graphics processing unit (GPU) 604, a digital signal processor (DSP) 606, a neural processing unit (NPU) 608, a multimedia processing unit 610, and wireless connectivity components 612.

[0087]

[0094] An NPU, such as 608, is generally a specialized circuit configured to implement all the necessary control and computational logic to execute machine learning algorithms, such as algorithms for processing artificial neural networks (ANN), deep neural networks (DNN), random forests (RF), etc. An NPU may alternatively be referred to as a neural signal processor (NSP), a tensor processing unit (TPU), a neural network processor (NNP), an intelligence processing unit (IPU), a vision processing unit (VPU), or a graph processing unit.

[0088]

[0095] An NPU, such as 608, is configured to accelerate the execution of common machine learning tasks, such as image classification, machine translation, object detection, and various other predictive models. In some examples, multiple NPUs may be instantiated on a single chip, such as a system-on-chip (SoC), while in other examples, they may be part of a dedicated neural network accelerator.

[0089]

[0096] An NPU may be optimized for training or inference, or in some cases may be configured to balance performance between both. In an NPU capable of performing both training and inference, the two tasks may still generally be performed independently.

[0090]

[0097] NPUs designed to accelerate training are generally configured to accelerate the optimization of new models, which is a highly computationally intensive operation that involves inputting an existing dataset (often labeled or tagged), iterating over the dataset, and then adjusting model parameters such as weights and biases to improve model performance. Generally, optimization based on mispredictions involves backpropagating through layers of the model and determining gradients to reduce prediction errors.

[0091]

[0098] NPUs designed to accelerate inference are generally configured to operate on complete models. Thus, such NPUs may be configured to input new data and rapidly process the data through already trained models to produce model outputs (e.g., inferences).

[0092]

[0099] In one implementation, the NPU 608 is part of one or more of the CPU 602, the GPU 604, and / or the DSP 606.

[0093]

[0100] In some examples, the wireless connectivity component 612 may include subcomponents for, for example, third generation (3G) connectivity, fourth generation (4G) connectivity (e.g., 4G LTE), fifth generation connectivity (e.g., 5G or NR), Wi-Fi connectivity, Bluetooth connectivity, and other wireless data transmission standards. The wireless connectivity processing component 612 is further connected to one or more antennas 614.

[0094]

[0101] The processing system 600 may also include one or more sensor processing units 616 associated with any type of sensor, one or more image signal processors (ISPs) 618 associated with any type of image sensor, and / or a navigation processor 620, which may include satellite-based positioning system components (e.g., GPS or GLONASS), as well as inertial positioning system components.

[0095]

[0102] The processing system 600 may also include one or more input and / or output devices 622, such as a screen, a touch-sensitive surface (including a touch-sensitive display), physical buttons, a speaker, a microphone, or the like.

[0096]

[0103] In some examples, one or more of the processors of processing system 600 may be based on the ARM or RISC-V instruction set.

[0097]

[0104] Processing system 600 also includes memory 624, which represents one or more static and / or dynamic memories, such as dynamic random access memory, flash-based static memory, etc. In this example, memory 624 includes computer-executable components that may be executed by one or more of the above-mentioned processors of processing system 600.

[0098]

[0105] In particular, in this example, memory 624 includes a measurement component 624A, an adaptive tuning component 624B, an estimation component 624C, a receiving component 624D, a transmitting component 624E, and a settings database 624F. The components shown and other components not shown can be configured to perform various aspects of the methods described herein.

[0099]

[0106] The processing system 600 further comprises an aperture tuner 626, for example as described above with respect to FIG.

[0100]

[0107] The processing system 600 further comprises an impedance tuner 628 as described above with respect to FIG.

[0101]

[0108] In general, the processing system 600 and / or its components may be configured to perform the methods described herein.

[0102]

[0109] Notably, aspects of the processing system 600 may be omitted in other cases. For example, the multimedia component 610, the ISP 618, and / or the navigation component 620 may be omitted in other aspects. Additionally, aspects of the processing system 600 may be distributed among multiple devices.

[0103] Alternatives

[0110] The following numbered clauses list alternative embodiments.

[0104]

[0111] Clause 1: receiving one or more operational characteristics of a wireless data transmission system of a device; and determining a target wireless data transmission system configuration based on the one or more operational characteristics using a wireless data transmission system configuration model; and implementing the target wireless data transmission system configuration in the wireless data transmission system.

[0105]

[0112] Clause 2: The method of clause 1, further comprising: determining, using a first estimator model, one or more estimated operating characteristics of the wireless data transmission system using one or more virtual wireless data transmission system configurations different from the target wireless data transmission system configuration; determining, using a second estimator model, one or more estimated performance metrics of the wireless data transmission system based on the one or more estimated operating characteristics; determining an actual performance metric based on the implemented target wireless data transmission system configuration; determining a best wireless data transmission system configuration among the target wireless data transmission system configuration and the one or more virtual wireless data transmission system configurations based on the actual performance metric and the one or more estimated performance metrics; and generating feedback data based on the determined best wireless data transmission system configuration.

[0106]

[0113] Clause 3: The method of clause 2, further comprising updating a wireless data transmission system configuration model based on the feedback data.

[0107]

[0114] Clause 4: The method of clause 3, wherein the wireless data transmission system configuration model is a reinforcement learning model.

[0108]

[0115] Clause 5: The method of clause 4, wherein the feedback data includes one of a reward signal for the reinforcement learning model when the target wireless data transmission system configuration matches the best wireless data transmission system configuration, or a penalty signal when the target wireless data transmission system configuration does not match the best wireless data transmission system configuration.

[0109]

[0116] Clause 6: The method of clause 4, wherein the feedback data includes a best wireless data transmission system configuration.

[0110]

[0117] Clause 7: The method of any one of clauses 2 to 6, wherein the actual performance metric includes an indicator provided by a network in wireless data communication with the device.

[0111]

[0118] Clause 8: The method of clause 7, wherein the indicator comprises at least one of a received signal received power (RSRP) metric, a received signal received quality (RSRQ) metric, a signal to noise metric (SNR / SINR), or a received signal strength indicator (RSSI).

[0112]

[0119] Clause 9: The method of any one of clauses 2 to 8, wherein the first estimator model comprises a neural network model.

[0113]

[0120] Clause 10: The method of any one of clauses 2 to 9, wherein the second estimator model comprises a neural network model.

[0114]

[0121] Clause 11: The method of any one of clauses 1 to 10, further comprising determining an optimal tuning code for the impedance tuner based on a target wireless data transmission system configuration.

[0115]

[0122] Clause 12: A method according to any one of clauses 1 to 11, wherein the one or more operating characteristics include a real impedance of an element of the wireless data transmission system, an imaginary impedance of an element of the wireless data transmission system, a frequency of the wireless data transmission system, and an antenna tuner state.

[0116]

[0123] Clause 13: The method of clause 12, wherein the element is an antenna of a wireless data transmission system and the target wireless data transmission system configuration includes an antenna tuner setting.

[0117]

[0124] Clause 14: The method of clause 13, wherein the antenna tuner settings include at least one of an aperture tuner setting or an impedance tuner setting.

[0118]

[0125] Clause 15: A processing system comprising a wireless data transmission system, a memory comprising computer-executable instructions, and one or more processors configured to execute the computer-executable instructions and cause the processing system to perform a method according to any one of clauses 1 to 14.

[0119]

[0126] Clause 16: A processing system comprising means for carrying out the method according to any one of clauses 1 to 14.

[0120]

[0127] Clause 17: A non-transitory computer-readable medium comprising computer-executable instructions that, when executed by one or more processors of a processing system, cause the processing system to perform the method of any one of clauses 1 to 14.

[0121]

[0128] Clause 18: A computer program product embodied on a computer-readable storage medium comprising code for performing the method according to any one of clauses 1 to 14.

[0122] Additional Considerations

[0129] The preceding description is provided to enable any person skilled in the art to practice the various aspects described herein. The embodiments described herein are not intended to limit the scope, applicability, or aspects described in the claims. Various modifications of these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects. For example, changes may be made in the function and arrangement of the elements described without departing from the scope of the disclosure. Various embodiments may omit, substitute, or add various procedures or components as appropriate. For example, the methods described may be performed in an order different from that described, and various steps may be added, omitted, or combined. Also, features described with respect to some embodiments may be combined in some other embodiments. For example, an apparatus may be implemented or a method may be practiced using any number of the aspects described herein. Additionally, the scope of the disclosure is intended to encompass such apparatus or methods practiced using other structures, functions, or structures and functions in addition to or other than the various aspects of the disclosure described herein. It should be understood that any aspect of the disclosure disclosed herein may be embodied by one or more elements of a claim.

[0123]

[0130] As used herein, the word "exemplary" means "serving as an example, instance, or illustration." Any aspect described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other aspects.

[0124]

[0131] As used herein, a phrase referring to "at least one of" a list of items refers to any combination of those items, including single members. By way of example, "at least one of a, b, or c" is intended to encompass a, b, c, ab, ac, bc, and abc, as well as any combination having multiples of the same element (e.g., aa, aaa, aab, aac, abb, acc, bb, bbb, bbc, cc, and ccc, or any other permutation of a, b, and c).

[0125]

[0132] As used herein, the term "determining" encompasses a wide variety of actions. For example, "determining" may include calculating, computing, processing, deriving, investigating, looking up (e.g., looking up in a table, database, or another data structure), ascertaining, etc. Also, "determining" can include receiving (e.g., receiving information), accessing (e.g., accessing data in a memory), etc. Also, "determining" can include resolving, selecting, choosing, establishing, etc.

[0126]

[0133] The methods disclosed herein include one or more steps or actions for achieving the method. The method steps and / or actions may be interchanged with one another without departing from the scope of the claims. In other words, unless a specific order of steps or actions is specified, the order and / or use of specific steps and / or actions may be modified without departing from the scope of the claims. Furthermore, various operations of the methods described above may be performed by any suitable means capable of performing the corresponding functions. The means may include various hardware and / or software component(s) and / or module(s), including but not limited to circuits, application specific integrated circuits (ASICs), or processors. In general, where there are operations illustrated in figures, those operations may have corresponding counterpart means-plus-function components with similar numbering.

[0127]

[0134] The following claims are not limited to the embodiments set forth herein, but are to be accorded the full scope consistent with the language of the claims. In the claims, reference to an element in the singular is not intended to mean "one and only one," but rather "one or more." Unless otherwise specified, the term "several" refers to one or more. No element of a claim is to be construed under the provisions of 35 U.S.C. 112(f) unless the element is expressly recited using the phrase "means of," or, in the case of a method claim, unless the element is recited using the phrase "step of." All structural and functional equivalents of the elements of the various embodiments described throughout this disclosure that are known or that later become known to those skilled in the art are expressly incorporated herein by reference and are intended to be encompassed by the claims. Moreover, nothing disclosed herein is intended to be made public, regardless of whether such disclosure is expressly recited in the claims.

Claims

1. receiving one or more operational characteristics of a wireless data transmission system of the device; determining a target wireless data transmission system configuration based on the one or more operating characteristics using a wireless data transmission system configuration model; determining, using a first estimator model, one or more estimated operating characteristics of the wireless data transmission system using one or more virtual wireless data transmission system configurations that differ from the target wireless data transmission system configuration; determining one or more estimated performance metrics of the wireless data transmission system based on the one or more estimated operating characteristics using a second estimator model; and determining actual performance metrics based on the implemented target wireless data transmission system configuration; determining a best wireless data transmission system configuration among the target wireless data transmission system configuration and the one or more virtual wireless data transmission system configurations based on the actual performance metric and the one or more estimated performance metrics; generating feedback data based on the determined best wireless data transmission system configuration; implementing the target wireless data transmission system configuration in the wireless data transmission system; A method comprising:

2. The method of claim 1 , further comprising: updating the wireless data transmission system configuration model based on the feedback data.

3. The method of claim 2 , wherein the wireless data transmission system configuration model is a reinforcement learning model.

4. The feedback data is a reward signal for the reinforcement learning model if the target wireless data transmission system configuration matches the best wireless data transmission system configuration; or The method of claim 3 , including one of a penalty signal if the target wireless data transmission system configuration does not match the best wireless data transmission system configuration.

5. The method of claim 3 , wherein the feedback data includes the best wireless data transmission system configuration.

6. The method of claim 1 , wherein the actual performance metric comprises an indicator provided by a network in wireless data communication with the device.

7. The indicator: Received Signal Received Power (RSRP) metric; Received Signal Reception Quality (RSRQ) metric; signal-to-noise metric (SNR / SINR), or The method of claim 6 , including at least one of a received signal strength indicator (RSSI).

8. The method of claim 1 , wherein the first estimator model comprises a neural network model.

9. The method of claim 1 , wherein the second estimator model comprises a neural network model.

10. The method of claim 1 , further comprising determining an optimal tuning code for an impedance tuner based on the target wireless data transmission system configuration.

11. The one or more operating characteristics are: a real impedance of an element of the wireless data transmission system; and an imaginary impedance of the element of the wireless data transmission system; a frequency of the wireless data transmission system; and an antenna tuner state.

12. the element is an antenna of the wireless data transmission system; The method of claim 11 , wherein the target wireless data transmission system configuration includes an antenna tuner setting.

13. The method of claim 12 , wherein the antenna tuner setting comprises at least one of an aperture tuner setting or an impedance tuner setting.

14. 1. An apparatus comprising: a wireless data transmission system; a memory comprising computer-executable instructions; one or more processors configured to execute the computer-executable instructions and cause the device to perform the steps of any one of claims 1 to 13; An apparatus comprising:

15. 14. A non-transitory computer-readable medium comprising computer-executable instructions that, when executed by one or more processors of a processing system, cause the processing system to perform the method of any one of claims 1 to 13.

Images