Patent Application
20210173011
This disclosure relates generally to integrated simulation and modeling based on artificial intelligence (AI).
Electronics is a field related to electrons. Electronic devices interact with electrons. An electronic device may influence an electron's behavior, and produce a technical effect. Some electronic devices manipulate electron behavior. For example, an amplifier may increase a signal's energy. Some electronic devices react to electron behavior. In an illustrative example, an attenuator may reduce a signal's energy. Various electronic devices may be active, passive, used alone, or configured in combination for use with other electronic devices.
Users of electronic devices include individuals, computer applications, organizations, government, and industry. Users may employ electronic devices to perform tasks the user may not otherwise accomplish without the electronic device. For example, a human mobile electronic device user may be able to join in a videoconference linking participants dispersed throughout the world. This facilitation may involve many complex devices interoperating through a network. Each electronic device in such a network of interoperating devices may include component, sub-system, or system level elements. For example, a single mobile communication device may include multiple computation, communication, and interface elements, configured together to perform the functions of the mobile communication device. In an illustrative example, each of the mobile communication device's component elements must operate together as designed, if the device is to function as intended.
Electronic device operating parameters include voltage, current, frequency, electromagnetic field strength, and other physical quantities. Device operating parameters may be determined based on device physics, the configuration of the device in a system, the device input, or other factors. A device behavioral model may predict device operating parameters. For example, a behavioral model may be designed to predict an electronic circuit's output determined as a function of an arbitrary input. In an illustrative example, predicting the operation of a complex device may require simulating models at component, sub-system, and system levels, to determine if the device might operate as designed before the device is physically constructed. A designer may expend significant time and effort simulating the operation of a system including many complex devices, each simulated based on multiple component, sub-system, and system models.
Apparatus and associated methods relate to augmenting a device model identified by artificial intelligence, with measurements of physical parameters, iteratively validating and verifying the augmented model until the augmented model satisfies a quality criterion determined as a function of the artificial intelligence, and automatically synthesizing an interactive simulation and measurement environment, based on the model. The model may be identified by the artificial intelligence based on measurement of a device operating characteristic. The physical parameter measurements the model is augmented with may be determined by the artificial intelligence, based on the model. The model may include a component, sub-system, and system model, permitting validation and verification through multiple levels. Various implementations may automatically generate a measurement scenario including communication commands configured to validate and verify the augmented model. Some designs may provide visualization of synthesized simulation and measurement output generated as a function of the validated and verified augmented model.
In an aspect, an apparatus is disclosed, comprising: a processor; and memory that is not a transitory propagating signal, said memory comprising instructions and data, and said memory further configured to be operably coupled to the processor, wherein the memory comprises encoded data and processor-executable program instructions, wherein the data and the instructions jointly configure and program the apparatus such that, when executed by the processor, the data and the instructions cause the apparatus to perform operations comprising: identifying the type of a device based on measuring a device operating characteristic; select a device behavioral model based on the identified device type, wherein the model is configured to model a parameter of the device; augment the model with a physical measurement of the modeled parameter identified as a function of the selected model; iteratively validate and verify the modeled parameter and the measured parameter, until an evaluation of the modeled parameter and the measured parameter satisfies a quality criterion determined as a function of an artificial intelligence; and provide access to the validated and verified model augmented with the measured physical parameter, for generating synthesized simulation and measurement output, based on the model.
The model may further comprise a physics-based model.
The operations performed by the apparatus may further comprise train the artificial intelligence with a physical model based on simulated data.
The model may further comprise a component model.
The model may further comprise a sub-system model.
The model may further comprise a system model.
Validate the modeled parameter and the measured parameter may further comprise determine if the physical parameter modeled is correct, based on measurement.
Verify the modeled parameter and the measured parameter may further comprise the measured parameter evaluated as a function of another verified measurement.
The quality criterion may be a threshold predetermined as a function of the modeled parameter.
The quality criterion may be a threshold predetermined as a function of the measured parameter.
The quality criterion may be a statistical function of the measured parameter.
The quality criterion may be a statistical function of the modeled parameter.
The quality criterion may be a function of the measured parameter evaluated in the frequency domain.
The quality criterion may be a function of the modeled parameter evaluated in the frequency domain.
The quality criterion may be a function of the measured parameter evaluated in the time domain.
The quality criterion may be a function of the modeled parameter evaluated in the time domain.
The quality criterion may be a function of bandwidth.
The quality criterion may be a function of frequency selectivity.
The quality criterion may be a function of sensitivity.
Provide access to the validated and verified model augmented with the measured physical parameter may further comprise the access provided to a graphical user interface configured to visualize the synthesized simulation and measurement output.
In another aspect, an apparatus is disclosed, comprising: a processor; and memory that is not a transitory propagating signal, said memory comprising instructions and data, and said memory further configured to be operably coupled to the processor, wherein the memory comprises encoded data and processor-executable program instructions, wherein the data and the instructions jointly configure and program the apparatus such that, when executed by the processor, the data and the instructions cause the apparatus to perform operations comprising: train an artificial intelligence with a physical model based on simulated data; identifying the type of a device under test (DUT), based on a measured device operating characteristic evaluated by the artificial intelligence; select a physics-based device behavioral model based on the identified device type, wherein the model is configured to predict a plurality of device parameters; augment the model with physical measurements of the modeled parameters, wherein the parameters the model is augmented with are identified by the artificial intelligence as a function of the selected model; iteratively validate and verify the modeled parameters and the measured parameters, until an evaluation of the modeled parameters and the measured parameters satisfies a quality criterion determined as a function of the artificial intelligence; and provide access in a graphical user interface to the validated and verified model augmented with the measured physical parameters, for generating a visualization of synthesized simulation and measurement output, based on the model.
The physics-based device behavioral model may further comprise: a component model; a sub-system model determined as a function of the component model; a system model determined as a function of the sub-system model; and a measurement model determined as a function of measurement setup.
Iteratively validate and verify the modeled parameters and the measured parameters may further comprise validate and verify based on model levels comprising: measurement, component, sub-system, and system, until the criterion is satisfied for all levels.
The measured parameter may be selected from the group consisting of current, electromagnetic field strength, frequency, impedance, and voltage.
The device under test may further comprise an amplifier.
The device under test may further comprise an attenuator.
The measured device operating characteristic may further comprise two-port insertion loss.
The measured device operating characteristic may further comprise two-port insertion gain.
In another aspect, an apparatus is disclosed, comprising: a processor; and memory that is not a transitory propagating signal, said memory comprising instructions and data, and said memory further configured to be operably coupled to the processor, wherein the memory comprises encoded data and processor-executable program instructions, wherein the data and the instructions jointly configure and program the apparatus such that, when executed by the processor, the data and the instructions cause the apparatus to perform operations comprising: train an artificial intelligence with a physical model based on simulated data; identifying the type of a device under test, based on measured device circuit network parameter evaluated by the artificial intelligence; select a physics-based device behavioral model based on the identified device type, wherein the model is configured to predict a plurality of device parameters, and wherein the model comprises: a component model; a sub-system model determined as a function of the component model; and a system model determined as a function of the sub-system model; augment the model with physical measurements of the modeled parameters, wherein the parameters the model is augmented with are identified by the artificial intelligence as a function of the selected model, and wherein the measured parameter is selected from the group consisting of current, electromagnetic field strength, frequency, impedance, and voltage; iteratively validate and verify the modeled parameters and the measured parameters based on a measurement scenario automatically prepared by the trained artificial intelligence, until an evaluation of the modeled parameters and the measured parameters for the component, sub-system, and system models satisfies a quality criterion determined as a function of the artificial intelligence; and provide access in a graphical user interface to the validated and verified model augmented with the measured physical parameters, for generating a visualization of synthesized simulation and measurement output, based on the model.
The operations performed by the apparatus may further comprise in response to determining a discrepancy between measured and modeled parameters, automatically adjust the measurement scenario.
The measurement scenario may further comprise commands configured to communicate with measurement instruments.
The artificial intelligence may be selected from the group consisting of a machine learning algorithm, an artificial neural network, and principle component analysis.
In another aspect, an apparatus is disclosed, comprising: a processor; and memory that is not a transitory propagating signal, said memory comprising instructions and data, and said memory further configured to be operably coupled to the processor, wherein the memory comprises encoded data and processor-executable program instructions, wherein the data and the instructions jointly configure and program the apparatus such that, when executed by the processor, the data and the instructions cause the apparatus to perform operations comprising: train an artificial intelligence with a physics-based behavioral model of a circuit based on simulated data generated based on the circuit model; identifying the type of a device under test as a circuit, based on measured circuit s-parameters evaluated as a function of frequency by the trained artificial intelligence; select a physics-based system-level behavioral model configured to predict a plurality of device, sub-system, and system parameters, wherein the model comprises: a circuit model; a sub-system model determined as a function of the circuit model; and a system model determined as a function of the sub-system model; augment the system-level model with physical measurements of the modeled parameters, wherein the parameters the system-level model is augmented with are identified by the artificial intelligence as a function of the circuit model, and wherein the measured parameter is selected from the group consisting of current, electromagnetic field strength, frequency, impedance, and voltage; iteratively validate and verify the modeled parameters and the measured parameters based on a measurement scenario automatically prepared by the trained artificial intelligence, until an evaluation of the modeled parameters and the measured parameters for the component, sub-system, and system models satisfies a quality criterion determined as a function of the artificial intelligence; and provide access in a graphical user interface to the validated and verified model augmented with the measured physical parameters, for generating a visualization of synthesized simulation and measurement output, based on the model.
The modeled parameters may further comprise an s-parameter.
The measured parameters may further comprise an s-parameter.
Augment the system-level model may further comprise link a physical measurement with a modeled parameter.
Provide access in the graphical user interface to the validated and verified model may further comprise a physical measurement correlated with a modeled parameter in a virtual environment.
The measurement scenario may further comprise commands configured to communicate with measurement instruments.
The measurement scenario may further comprise physical parameters measured by a vector network analyzer.
The operations performed by the apparatus may further comprise exchange a measured parameter with a simulation tool.
The operations performed by the apparatus may further comprise exchange a modeled parameter with a simulation tool.
The operations performed by the apparatus may further comprise exchange data with a Computer Aided Design (CAD) environment or a Printed Circuit Board (PCB) layout environment, to overlap with measurement or simulation data.
The present disclosure teaches a Simulation and Measurement System or Method. The Simulation and Measurement System may be a computer-implemented Simulation and Measurement Platform. The Simulation and Measurement Method may be a process implementing Simulation and Measurement Platform features. The computer-implemented Simulation and Measurement Platform may include computer hardware. The computer-implemented Simulation and Measurement Platform may include computer software. The computer hardware may include a processor and a memory. The memory may encode processor executable program instructions configured to cause the hardware to perform the disclosed Simulation and Measurement operations. The computer hardware may include interfaces designed to permit the processor to interact with and capture measurements from a device under test using various test and measurement instruments.
An exemplary Simulation and Measurement Platform implementation of a software and hardware integrated platform for a physics-based artificial intelligence (AI) configured to combine measurement and simulation may permit a user to acquire data from measurement instruments, process the data, and synthesize the measured and simulated data in a single environment. The disclosed software includes implementation of the process using computer codes, including processor executable program instructions, which may be executed on a personal computer, server, digital signal processor, cloud computing, or other computational hardware platform as may be available to one of ordinary skill. The hardware may include an interface configured to interact with or control a measurement instrument, an electrical or electronics component, the device under test, or any other physical component used in the measurement process. The measurement instrument may be a hardware apparatus configured to acquire a physical quantity and convert the physical quantity to a digital or analog signal which can communicate with the software platform. Some examples of the measurement instruments include sensors, data acquisition cards (DAQ), spectrum analyzers, oscilloscopes, vector network analyzers (VNA), time domain reflectometers (TDR), signal analyzers, and the like, as would be known to one of ordinary skill.
An exemplary Simulation and Measurement Platform may be configured to provide an integrated software and hardware environment facilitating the data measurement, simulation, visualization, correlation, and management for these design stages, with a cohesive integrated software and hardware platform capable of communicating and interacting with various devices and instruments, including, for example: measurement instruments; control electronic boards (for example, data acquisition (DAQ), analog to digital converter (ADC)/digital to analog converter (DAC), single-board computer, programmable logic controller (PLC), relay, and the like); control motorized moving structures or robotic arms; and, exchange data with simulation tools, computer-aided design (CAD), and printed circuit board (PCB) layout tools.
An exemplary Simulation and Measurement Platform software implementation may include a physics-based environment configured to perform operations such as: acquiring physical quantities based on pre-defined procedures/standards or customized procedures; self-correlating or cross-correlating various physical quantities; 0D, 1D, 2D, or 3D spatial measurement in time, frequency, or time-frequency domain; exchange data with simulation tools; post-process acquired data from measurement or simulation environments; manage large quantities of data across multiple users and over a communication network; simulate a virtual electromagnetic or circuit equivalent environment; exchange data with CAD and PCB layout environments to overlap with measurement or simulation data; provide access to a library of measurement components configured with a behavioral model designed to simulate component characteristics; and, provide an intuitive visualization and reporting environment.
An exemplary Simulation and Measurement Platform software design may be configured to facilitate an iterative validation and verification process, to provide an integrated measurement and simulation environment. The software may be configured to communicate with measurement instruments, and standard machine and/or human readable data exchange formats, such as Touchstone (SnP), IEC TR 91967-1-1, Measurement Data Format (MDF), and other formats as may be available to one of ordinary skill. The software may be configured to verify imported data in the imported environment for accuracy and compatibility with physical quantities. In a more advanced scenario, the software may be configured to permit automatically or semi-automatically adjusting simulation setup or measurement instrument settings, if discrepancies are observed. For example, the software may be configured to change meshing criteria accordingly, if the simulation tool detects a specific or predetermined pattern from measured near-field scanned data. In another example scenario, if a measurement instrument determines based on simulation results that the expected signal level is smaller than the current noise level, the measurement bandwidth or dynamic range may be adjusted accordingly. An exemplary Simulation and Measurement Platform software implementation may be configured in a modular design, to have the capability to define a measurement setup based on user-defined modules. In an illustrative example, such user-defined modules may be assembled together as templates for measurement standards, or common practices. The AI may also be implemented as library of AIs, and may be controlled by a higher level AI.
An exemplary Simulation and Measurement Platform simulation design may include implementation of a physical model in the disclosed software platform, or in a third-party software platform which can exchange data with the software platform. The physics-based feature of the software may be implemented by software configured to process the data based on the electromagnetics, physical interpretation of the quantity, the physical dimensions (units), or mathematical models representing the physics governing the device or system under test. In an illustrative example, the software may be configured to use artificial intelligence models to create an augmented physical environment (APE) through the graphical user interface (GUI) for easy correlation of physical measured data with simulated data in a virtual environment in the software platform. The virtual environment includes an implementation in the software platform representing the physical model of the hardware platform. The augmented physical environment includes linking and correlation between components of the physical quantity or the physical device under test (DUT) to the physics-based model in the virtual environment. This integration is designed to perform the data-to-information conversion even for a user without advanced training. The data may include the raw numerical data acquired by a measurement instrument. The information includes presentation of the data in a format that is easy to interpret for a human user, such as, for example, plots, diagrams, flow-charts, and the like. The disclosed software implementation facilitates training of an artificial intelligence with simulated data generated by a physics-based model. The artificial intelligence may be implemented as a machine learning (ML) algorithm. The artificial intelligence may be configured with optimizations such as, for example, artificial neural networks (ANN), embedded mapping, or principle component analysis (PCA). The trained AI may then be used to automatically prepare a measurement scenario. The measurement scenario may include a set of communication commands (for example, SCIPI) configured to communicate with measurement instruments.
An exemplary Simulation and Measurement Platform may perform systematic synthesis of measurement and simulation based on iterative validation and verification. Validation may include the software performing operations including measurement and simulation to answer the question “Is the software measuring and/or modeling the right physical quantity?” to assure the design of each measurement or simulation process and the choice of metric meets the final needs, in manufacturing, test, and/or operational conditions. Verification may include the software performing operations including measurement and simulation to answer the question “Is the software measuring and/or modeling the physical quantity correctly?” based on evaluating the measurement process with another verified measurement or simulation, and evaluating the simulation design process by another verified simulation or measurement with shared governing physics and shared evaluation metrics.
In an illustrative example, the disclosed iterative validation and verification process may be based on evaluation of each contributing component to a system level measurement and simulation. The software may then generate (or acquire from a third-party software) a physics-based behavioral model which represents the relevant electrical or electromagnetic behavior of the components or sub-systems under test and the relationship between components and the system level evaluation. Once individual components of the simulation and measurement processes are verified, the components can be used as a tool for understanding the system behavior which are otherwise difficult to characterize.
Various Simulation and Measurement Platforms may achieve one or more technical effect. For example, some Simulation and Measurement Platforms may improve a user's ease of access to system simulation. This facilitation may be a result of reducing the user's effort adjusting a device under test model and configuring the device model with a system model in the user's simulation and measurement setup. In some Simulation and Measurement Platforms, a device under test model and the device parameters to be simulated may be automatically selected for a user based on a device physical measurement evaluated by artificial intelligence. For example, the artificial intelligence may be trained with simulated data generated from a physics-based device model, and the trained artificial intelligence may identify a device model based on matching measurement of a physical device under test with the simulated data. Such automatic device under test and simulation parameter identification may reduce a user's exposure to the risk of error in model selection, and improve the user's confidence that a simulated model type matches the device under test.
Some Simulation and Measurement Platforms may reduce a user's effort obtaining verified measurement results related to a user's testing or development of an electronic system design. Such reduced effort obtaining verified measurement results may be a result of an iterative validation and verification process evaluating each contributing component to a system level measurement and simulation. Such verified measurement results may improve the user's simulation, modeling, and measurement experience. For example, verified individual components of the simulation and measurement processes may aid evaluation of complex system behavior, permitting a user to adjust system designs more quickly, and improve the accuracy or usefulness of research and development testing. In some Simulation and Measurement Platforms, a user's understanding of system behavior may be improved. Such improved understanding of system behavior may be a result of providing the user access to a visual Augmented Physical Environment (APE) correlating device under test physical measurement with simulated data in a virtual lab. For example, a user may more quickly understand the effect of a design change, based on augmented simulation visualized by the APE linking and correlating device under test physical measurements with data generated by a physics-based model in a measurement scenario generated by artificial intelligence.
In an illustrative example, design and production of an electronic board may require multiple iterations of simulation, measurement, validation, and verification, as well as pre-compliance and compliance tests. An exemplary implementation of a software and hardware integrated platform for a physics-based artificial intelligence (AI) configured to combine measurement and simulation is herein disclosed. The exemplary software implementation creates a cohesive and integrated platform designed to combine measurement and simulation, based on artificial intelligence models.
A “working example” of one aspect/embodiment of the instant invention works as follows for, inter alia, Cable Impedance Measurement and Simulation:
In one such example, the instant invention utilizes its “V-model” (
Here, the “system” would be the device being tested, here an HDMI cable over the ground plane, e.g. a wire with diameter 4 mm, substrate height 32 mm, and substrate dielectric Er 1.1; Output impedance 206 Ohms.
Here, the “sub system” is the cable assembly over the ground plane.
Here, the “component” would be, e.g., ferrite.
Here, another “component” would be the ground plane.
This example of the invention at work would have an initial Measurement Setup comprising a measuring instrument, here a vector network analyzer, and the device tested would be the HDMI cable over ground plane.
The invention then performs sub-system modeling of the cable over ground plane using third-party impedance calculator.
The simulation modeling of the system, sub-system, and components are performed in a third-party circuit simulator.
The model of the transfer impedance of the cable is extracted from the impedance calculator, which could be, for example, 206 Ohms with, for example, a time delay of 0.4 nano-seconds.
The invention then performs system, sub-system, and component level modeling using said third-party software tool.
The model of ferrite and the parameters of the model are adjusted iteratively using syntheses of measurement and simulation per the invention's apparatus and method.
Results: Measurement results are “verified” in the measurement setup with or without the ferrite on the cable assembly. Namely, the invention is therein able to compare measurements of DUT with ferrite and without ferrite.
The invention is then able to measure return loss of the cable using a vector network analyzer. Return loss can then be measured showing the various margins (e.g. between 10 and 60 dBΩ) at various frequencies (between, e.g. 1 MHz and 1 GHz).
Based on the above synthesis of the measurement and simulation, the model parameters of the ferrite on the cable assembly would be, for example, extracted as follows:
In this example, the above parameters are then visually modeled and displayed via Graphical User Interface appropriately configured to visually illustrate the above parameters in their assiciated units and ratios.
The details of various aspects are set forth in the accompanying drawings and the description below. Other features and advantages will be apparent from the description and drawings, and from the claims.
Like reference symbols in the various drawings indicate like elements.
To aid understanding, this document is organized as follows. First, synthesizing an integrated simulation and measurement environment is briefly introduced with reference to
In an illustrative example, the AIISMP 125 may retrieve the DUT 110 model 140 from the model database server 130. The model 140 may be a behavioral model configured to predict a DUT 110 physical operating parameter. The behavioral model 140 may be a physics-based model of a component, sub-system, or system. The AIISMP 125 may augment the model 140 with measurement 145 of parameter 150 to create the synthesized integrated simulation and measurement environment 155. The AIISMP 125 may provide the synthesized integrated simulation and measurement environment 155 to the user 105 via the user device 120 user interface. The synthesized integrated simulation and measurement environment 155 may be referred to as an APE (Augmented Physical Environment).
In the depicted implementation, the AIISMP 125 generation of the DUT 110 APE begins at step 160 with the AIISMP 125 selecting the physics-based DUT 110 model 140 and physical parameters 150 to be validated and verified. The model 140 may include setup, component, sub-system, and system model levels. At step 165, the AIISMP 125 captures physical measurement 145 from the DUT 110 using the measurement instrument 135. The AIISMP 125 compares the measurement 145 to the model 140 prediction of the parameter 150. At step 170, the AIISMP 125 validates the selected parameter 150, based on the comparison. At step 175, the AIISMP 125 verifies the modeled parameter 150 based on the measurement 145. At step 180, the AIISMP 125 performs a test to determine if the parameter 150 has been validated and verified with the measurement 145. Upon a determination by the AIISMP 125 at step 180 the parameter 150 has not been validated and verified, the AIISMP 125 continues at step 190 augmenting the model 140 with the measurement 145, and the AIISMP 125 operation continues at step 165. Upon a determination by the AIISMP 125 at step 180 the parameter 150 has been validated and verified, the AIISMP 125 at step 185 performs a test to determine if all parameters have been validated and verified for all model 140 levels. Upon a determination by the AIISMP 125 at step 185 all parameters 150 have not been validated and verified for all model 140 levels, the AIISMP 125 continues at step 190 augmenting the model 140 with the measurement 145. Upon a determination by the AIISMP 125 at step 185 all parameters 150 have been validated and verified for all model 140 levels, the AIISMP 125 at step 195 provides the validated and verified augmented physical environment 155 to the user device 120. The process may repeat.
In the depicted implementation, the process V-model 500 includes the repetition of each step to fulfill the evaluation of validation and verification processes. In the illustrated implementation, the evaluation of validation and verification based on the behavioral model 140 is validated and verified based on shared physics and measurements 150 captured by the measurement instrument 135 from the device under test 110. In the depicted implementation, the evaluation of validation and verification based on shared metrics is repeated for the component simulation 505a, the sub-system simulation 505b, and the system simulation 505c until comparable results are obtained. The determination that comparable results have been obtained between the system measurement and system simulation may be based on a quality criterion evaluated by an artificial intelligence. Additionally, verified measurement results can be used as inputs to simulation tools, to increase the accuracy and efficiency of the simulation process. In this process V-model 500 combining the measurement and simulation, the behavioral model 140 in simulation may be physics-based and share the same governing physics as the measurement setup. As shown in
The depicted method 800 begins at step 805 with the processor 305 performing a test to determine if the device under test is an attenuator or an amplifier. The determination may be based on measurement data evaluated as a function of an AI trained with simulated data generated by a model of a known device type.
Upon a determination by the processor 305 at step 805 the device under test is an attenuator, the method continues at step 810 with the processor 305 selecting an attenuator simulation model, and the method continues at step 820.
Upon a determination by the processor 305 at step 805 the device under test is an amplifier, the method continues at step 815 with the processor 305 selecting an amplifier simulation model, and the method continues at step 820.
At step 820 the processor 305 activates the trained AI to govern the Simulation and Measurement Platform measurement scenario based on modeled and measured parameters, and the method continues at step 825.
At step 825, the processor 305 receives from the AI hard limits for modeled and measured parameters determined by the AI.
At step 830, the processor 305 sets measurement instrument parameters. In this example scenario, the software communicates with the VNA using a communication protocol such as Standard Commands for Programmable Instruments (SCPI). In some implementations, the user may set in the software that the parameter of interest is an S-parameter and the measurement instrument is a VNA. In this example, the AI configures the initial parameters in the software platform to transmit a pilot signal from port 1 and receive a signal from port 2 of the VNA. The measurement instrument parameters set by the processor 305 may be based on the hard parameter limits received by the processor 305 from the AI. The software may then be reconfigured according to the new information, with the capability of user interaction to change the parameters. This process may be implemented by machine learning algorithms such as an active learning model. The measurement instrument parameters set by the processor 305 may include, for example, Start frequency, Stop frequency, IF bandwidth, Power, Number of points, S-parameter, Sweep type, or other parameters as may be known to one of ordinary skill.
At step 835, the processor 305 reads data from the instrument. The data read by the processor 305 from the instrument may be measurement data captured from the device under test. The processor 305 activates the AI to analyze the data read from the instrument. After the initial parameters on the instrument are set, the software reads the S-parameter of the DUT. Another layer of AI performs an analysis on the acquired data. This analysis is performed to confirm the measured S-parameter follows the expected values based on the simulated physical model. In this example, the AI analyzes the transmitted and received pilot signal to understand the physical property of the DUT, and match it to a physical component based on the simulated parameters of the DUT. In this scenario the simulation model of the attenuator is the mathematical attenuation factor of the magnitude of a sinusoidal signal on the output of the attenuator compared to the input of the attenuator. In the case of the amplifier, the simulation model is the mathematical amplification factor of the magnitude of a sinusoidal signal on the output of the amplifier compared to the input of the amplifier. In this case, the AI provides a probabilistic prediction of the type of the DUT and the probabilistic prediction of the parameters of the DUT such as gain or loss. The AI provides a confidence level for the predicted parameters based on the pilot signal.
At step 840, the processor 305 performs a test to determine if the data read from the instrument by the processor 305 at step 835 matches the simulation model selected based on the determination by the processor 305 at step 805. The model may be an attenuator or amplifier simulation model. Upon a determination by the processor 305 at step 840 the data read from the instrument matches the selected model, the method continues at step 845. Upon a determination by the processor 305 at step 840 the data read from the instrument does not match the selected model, the processor 305 activates the AI to analyze the data, and the method continues at step 850.
At step 845, the processor 305 reads and saves the data read from the instrument, plots the data per the user's configuration, and the method ends.
At step 850, the processor 305 performs a test to determine if a physical test setup issue has been detected, based on the AI data analysis performed by the processor 305 at step 840. The AI may decide if the mismatch is due to improper parameters on the instruments, or due to an issue in the physical measurement setup. In the case of improper parameters, the AI may reconfigure the settings in the instrument, and repeat the data acquisition process described above. Upon a determination by the processor 305 at step 850 that a physical test setup issue has been detected, the method continues at step 860. Upon a determination by the processor 305 at step 850 that a physical test setup issue has not been detected, the method continues at step 855.
At step 855, the processor 305 activates the AI to analyze the data read from the instrument, and the processor 305 changes the instrument parameters based on the AI data analysis. The method continues at step 830.
At step 860, the processor 305 indicates the user needs to change the physical setup. The processor 305 may notify the user to make proper modification in the setup, and the software may repeat the data acquisition process described above. The method continues at step 830.
In some implementations, the method may repeat. In various implementations, the method may end.
The reference numbers and their respective elements depicted by the Drawings are summarized as follows.
Although various features have been described with reference to the Drawings, other features are possible.
In the present disclosure, various features may be described as being optional, for example, through the use of the verb “may.” For the sake of brevity and legibility, the present disclosure does not explicitly recite each and every permutation that may be obtained by choosing from the set of optional features. However, the present disclosure is to be interpreted as explicitly disclosing all such permutations. For example, a system described as having three optional features may be implemented in seven different ways, namely with just one of the three possible features, with any two of the three possible features or with all three of the three possible features. The respective implementation features, even those disclosed solely in combination with other implementation features, may be combined in any configuration excepting those readily apparent to the person skilled in the art as nonsensical.
A number of implementations have been described. Nevertheless, it will be understood that various modifications may be made. For example, the steps of the disclosed techniques may be performed in a different sequence, components of the disclosed systems may be combined in a different manner, or the components may be supplemented with other components. Accordingly, other implementations are contemplated, within the scope of the following claims.
This application claims the benefit of U.S. Provisional Application No. 62/945,008, titled “Physics-Based Artificial Intelligence Integrated Simulation and Measurement Platform,” filed by Hamed Kajbaf, on 6 Dec. 2019. This application incorporates the entire contents of the above-referenced application herein by reference.
| Number | Date | Country | |
|---|---|---|---|
| 62945008 | Dec 2019 | US |
