Integrated circuit electromagnetic sensitive area positioning method, device, equipment and system

Abstract

The application relates to an integrated circuit electromagnetic sensitive area positioning method, device, equipment and system. The method comprises the following steps: obtaining a spatial coordinate grid corresponding to a target integrated circuit and a target experiment condition; inputting the target experiment condition of each coordinate point in the spatial coordinate grid into a preset fault prediction model to obtain a sensitivity matrix corresponding to the spatial coordinate grid, the fault prediction model being obtained by training historical electromagnetic fault injection experiment data, the experiment data comprising historical experiment conditions and result data, the historical experiment conditions comprising historical pulse power, historical pulse time delay and historical injection position coordinates; and generating a sensitivity heat map corresponding to the target integrated circuit according to the sensitivity matrix. The method can intuitively, clearly and comprehensively position the electromagnetic sensitive area on the surface of the target integrated circuit.

Description

Technical Field

[0001] This application relates to the field of integrated circuit testing technology, and in particular to a method, apparatus, equipment and system for locating electromagnetically sensitive areas of integrated circuits. Background Technology

[0002] In order to perform functional safety testing, security assessment and design effective protection measures for MCUs (Micro Controller Units) to deal with attack methods such as electromagnetic fault injection (EMFI), it is necessary to locate the electromagnetically sensitive areas of the MCU integrated circuit.

[0003] Traditional techniques rely on specific criteria, such as analyzing electromagnetic radiation to narrow down the search area in advance; or, they apply automated heuristic search algorithms to optimize EMFI parameters by simulating biological evolution to find the optimal parameter combination, thereby determining the electromagnetically sensitive region.

[0004] However, the above positioning method has the problem of incomplete positioning of electromagnetically sensitive areas of integrated circuits. Summary of the Invention

[0005] Therefore, it is necessary to provide a method, apparatus, device, and system for locating electromagnetically sensitive areas of integrated circuits that can perform comprehensive positioning, addressing the aforementioned technical problems.

[0006] In a first aspect, this application provides a method for locating electromagnetically sensitive areas of an integrated circuit, including:

[0007] Obtain the spatial coordinate grid and experimental conditions corresponding to the target integrated circuit;

[0008] Input the target experimental conditions of each coordinate point in the spatial coordinate grid into the preset fault prediction model to obtain the sensitivity matrix corresponding to the spatial coordinate grid. The fault prediction model is trained based on the experimental data in the historical electromagnetic fault injection experiment process. The experimental data includes historical experimental conditions and result data. The historical experimental conditions include historical pulse power, historical pulse delay and historical injection position coordinates.

[0009] Based on the sensitivity matrix, a sensitivity heatmap corresponding to the target integrated circuit is generated.

[0010] In one embodiment, the target experimental conditions include target pulse power, target pulse delay, and target injection location coordinates;

[0011] Input the calibration conditions of each coordinate point in the spatial coordinate grid into the preset fault prediction model, including:

[0012] Based on the target injection location coordinates, target pulse power, and target pulse delay, generate parameter vectors corresponding to each coordinate point;

[0013] The parameter vector is subjected to feature enhancement processing to obtain the feature vector;

[0014] Input the feature vector into the fault prediction model to obtain the sensitivity corresponding to the coordinate point;

[0015] A sensitivity matrix is ​​generated based on the sensitivity of each coordinate point.

[0016] In one embodiment, the feature vector includes interactive features, multinomial features, and nonlinear features;

[0017] The parameter vector is subjected to feature enhancement processing to obtain the feature vector, including:

[0018] At least two features from the target injection position coordinates, target pulse power and target pulse delay in the parameter vector are combined, and multiplication is performed in each combination to obtain multiple interactive features.

[0019] By performing exponential operations on the target injection position coordinates, target pulse power, and target pulse delay in the parameter vector, multiple polynomial features are obtained.

[0020] By using preset nonlinear functions to transform the target injection position coordinates, target pulse power, and target pulse delay, multiple nonlinear characteristics are obtained.

[0021] In one embodiment, the method further includes:

[0022] Adjust the target experimental conditions, return to the step of inputting the target experimental conditions of each coordinate point in the spatial coordinate grid into the preset fault prediction model, and obtain sensitivity heatmaps under multiple different target experimental conditions;

[0023] A comprehensive sensitivity map is obtained by averaging multiple sensitivity heatmaps point by point.

[0024] In one embodiment, the method further includes:

[0025] The comprehensive sensitivity map is normalized to obtain the relative coordinate sensitivity map.

[0026] In one embodiment, the method further includes:

[0027] Obtain the initial prediction model and raw data;

[0028] The raw data is standardized to obtain experimental data;

[0029] The initial prediction model was trained using experimental data to obtain the fault prediction model.

[0030] Secondly, this application also provides an integrated circuit electromagnetic sensitive area positioning device, comprising:

[0031] The data acquisition module is used to acquire the spatial coordinate grid and experimental conditions corresponding to the target integrated circuit.

[0032] The matrix generation module is used to input the target experimental conditions of each coordinate point in the spatial coordinate grid into the preset fault prediction model to obtain the sensitivity matrix corresponding to the spatial coordinate grid. The fault prediction model is trained based on the experimental data in the historical electromagnetic fault injection experiment process. The experimental data includes historical experimental conditions and result data. The historical experimental conditions include historical pulse power, historical pulse delay and historical injection position coordinates.

[0033] The image generation module is used to generate a sensitivity heatmap corresponding to the target integrated circuit based on the sensitivity matrix.

[0034] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the integrated circuit electromagnetic sensitive area positioning method as described in the first aspect.

[0035] Fourthly, this application also provides an integrated circuit electromagnetic sensitive area positioning system, which includes a testing device and a computer device as described in the third aspect. The testing device is connected to the computer device and is used to perform electromagnetic fault injection experiments on the target integrated circuit, obtain experimental data, and send the experimental data to the computer device.

[0036] Fifthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the integrated circuit electromagnetic sensitive area location method as described in the first aspect.

[0037] The aforementioned method, apparatus, equipment, and system for locating electromagnetically sensitive areas of integrated circuits, by traversing the coordinate points on the spatial coordinate grid corresponding to the target integrated circuit and using a fault prediction model to predict the probability of failure success for each coordinate point under the target experimental conditions, ultimately generates a sensitivity heat map. This map intuitively represents the distribution of electromagnetically sensitive areas of the target integrated circuit under the target experimental conditions. Compared with existing technologies that search for discrete optimal parameter points, this method can intuitively, clearly, and comprehensively locate the electromagnetically sensitive areas on the surface of the target integrated circuit. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 This is an application environment diagram of an integrated circuit electromagnetic sensitive area localization method in one embodiment;

[0040] Figure 2 This is a flowchart illustrating a method for locating electromagnetically sensitive areas of an integrated circuit in one embodiment.

[0041] Figure 3 This is a flowchart illustrating the steps for obtaining the sensitivity corresponding to a coordinate point in one embodiment;

[0042] Figure 4 This is a schematic diagram of the structure of a fault prediction model in one embodiment;

[0043] Figure 5 This is a flowchart illustrating the method for locating electromagnetically sensitive areas of an integrated circuit in another embodiment;

[0044] Figure 6 Here is the ROC curve corresponding to the fault prediction model in one embodiment;

[0045] Figure 7(a) is a three-dimensional graph of the overall sensitivity corresponding to target experimental condition 1 in one embodiment;

[0046] Figure 7(b) is a two-dimensional graph of the overall sensitivity corresponding to target experimental condition 1 in one embodiment;

[0047] Figure 8(a) is a three-dimensional graph of the overall sensitivity corresponding to target experimental condition 2 in one embodiment;

[0048] Figure 8(b) is a two-dimensional graph of the overall sensitivity corresponding to target experimental condition 2 in one embodiment;

[0049] Figure 9(a) is a three-dimensional graph of the comprehensive sensitivity corresponding to target experimental condition 3 in one embodiment;

[0050] Figure 9(b) is a two-dimensional graph of the overall sensitivity corresponding to target experimental condition 3 in one embodiment;

[0051] Figure 10(a) is a three-dimensional graph of the overall sensitivity corresponding to target experimental condition 4 in one embodiment;

[0052] Figure 10(b) is a two-dimensional graph of the overall sensitivity corresponding to target experimental condition 4 in one embodiment;

[0053] Figure 11(a) is a three-dimensional plot of the relative coordinate sensitivity map in one embodiment;

[0054] Figure 11(b) is a two-dimensional plot of the relative coordinate sensitivity map in one embodiment;

[0055] Figure 12 This is a structural block diagram of an integrated circuit electromagnetic sensitive area positioning device in one embodiment;

[0056] Figure 13 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0058] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0059] The integrated circuit electromagnetic sensitive area location method provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104, or it can be located in the cloud or on another network server.

[0060] Terminal 102 is used to conduct a series of electromagnetic fault injection physical experiments on the target integrated circuit, obtain experimental data, and send it to server 104. Server 104 is used to obtain the spatial coordinate grid and target experimental conditions corresponding to the target integrated circuit; input the target experimental conditions of each coordinate point in the spatial coordinate grid into a preset fault prediction model to obtain the sensitivity matrix corresponding to the spatial coordinate grid. The fault prediction model is trained based on experimental data from historical electromagnetic fault injection experiments. The experimental data includes historical experimental conditions and result data. The historical experimental conditions include historical pulse power, historical pulse delay, and historical injection position coordinates; based on the sensitivity matrix, a sensitivity heatmap corresponding to the target integrated circuit is generated.

[0061] The terminal 102 can be, but is not limited to, various pulse generators, electromagnetic probes, and data acquisition devices. The server 104 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0062] In one exemplary embodiment, such as Figure 2 As shown, a method for locating electromagnetically sensitive areas of an integrated circuit is provided, which can be applied to... Figure 1 Taking server 104 as an example, the explanation includes the following steps 202 to 206. Wherein:

[0063] Step 202: Obtain the spatial coordinate grid and experimental conditions corresponding to the target integrated circuit.

[0064] Specifically, a spatial coordinate grid with a preset resolution is defined on the XY coordinate plane of the target integrated circuit. This spatial coordinate grid covers the entire or interested physical region of the target integrated circuit. For example, the resolution can be 100×100; a higher resolution can be chosen if computational conditions permit.

[0065] The target experimental conditions can be preset factors that are sensitive to electromagnetic fault injection, such as pulse intensity, pulse injection delay, and pulse waveform.

[0066] Step 204: Input the target experimental conditions of each coordinate point in the spatial coordinate grid into the preset fault prediction model to obtain the sensitivity matrix corresponding to the spatial coordinate grid.

[0067] The fault prediction model was trained based on experimental data from historical electromagnetic fault injection experiments. The experimental data included historical experimental conditions and results. The historical experimental conditions included historical pulse power P0, historical pulse delay D0, and historical injection position coordinates (X0, Y0).

[0068] The result data can be a Boolean value, used to indicate whether the corresponding experimental process successfully triggered the preset fault event. For example, a Boolean value of 1 indicates that the experimental process successfully triggered the preset fault event; a Boolean value of 0 indicates that the experimental process failed to trigger the preset fault event.

[0069] The fault prediction model can be a learning model based on Convolutional Neural Networks (CNN), Support Vector Machines (SVM), or Gradient Boosting Machines (GBM) to establish a nonlinear mapping relationship between the target experimental conditions of coordinate points and the probability of fault success, thereby obtaining the predicted probability value corresponding to the coordinate point, i.e., the sensitivity score. By traversing all coordinate points on the spatial coordinate grid, the sensitivity score of each coordinate point under the target experimental conditions is obtained, and then the sensitivity scores corresponding to each coordinate point are reorganized into a two-dimensional matrix, i.e., the sensitivity matrix.

[0070] Step 206: Generate a sensitivity heatmap corresponding to the target integrated circuit based on the sensitivity matrix.

[0071] In the process of generating the sensitivity heatmap, a color mapping scheme is set to define the correspondence between values ​​and colors, so as to intuitively display the high-sensitivity areas in the target integrated circuit.

[0072] In the aforementioned method for locating electromagnetically sensitive areas of integrated circuits, by traversing the coordinate points on the spatial coordinate grid corresponding to the target integrated circuit, a fault prediction model is used to predict the probability of failure success for each coordinate point under the target experimental conditions. Finally, a sensitivity heat map is generated, which intuitively represents the distribution of electromagnetically sensitive areas of the target integrated circuit under the target experimental conditions. Compared with the existing technology of finding discrete optimal parameter points, this method can intuitively, clearly, and comprehensively locate the electromagnetically sensitive areas on the surface of the target integrated circuit.

[0073] In one exemplary embodiment, based on Figure 2 The embodiment shown includes target pulse power, target pulse delay, and target injection position coordinates.

[0074] like Figure 3 As shown, the target experimental conditions for each coordinate point in the spatial coordinate grid are input into the preset fault prediction model, including:

[0075] Step 302: Generate parameter vectors corresponding to each coordinate point based on the target injection location coordinates, target pulse power, and target pulse delay.

[0076] The parameter vector is a vector containing the original features of four dimensions: target injection position coordinates (X1, Y1), target pulse power P1, and target pulse delay D1.

[0077] Step 304: Perform feature enhancement processing on the parameter vector to obtain the feature vector.

[0078] In one possible implementation, the feature vector includes interactive features, polynomial features, and nonlinear features. Step 304 may further include: combining at least two features from the target injection position coordinates, target pulse power, and target pulse delay in the parameter vector, and performing multiplication operations in each combination to obtain multiple interactive features; performing exponential operations on the target injection position coordinates, target pulse power, and target pulse delay in the parameter vector respectively to obtain multiple polynomial features; and transforming the target injection position coordinates, target pulse power, and target pulse delay using a preset nonlinear function to obtain multiple nonlinear features.

[0079] In some embodiments, the interaction features may include first-order interaction features obtained by multiplying any two original features, and higher-order composite interaction features obtained by multiplying three or more original features.

[0080] For example, first-order interaction features may include, but are not limited to: X1×Y1, Y1×P1, X1×D1, P1×D1; higher-order composite interaction features may include, but are not limited to: X1×Y1×P1; polynomial features may include, but are not limited to: X1 2 P1 3 D1 2 Nonlinear characteristics may include, but are not limited to, characteristics obtained by transforming the target pulse power and / or target pulse delay using nonlinear functions such as logarithmic or exponential functions.

[0081] In this embodiment, by combining and transforming the original four-dimensional features, feature vectors containing dozens or hundreds of dimensions can be obtained, so as to fully capture the more complex and multi-dimensional interactions between the original features, provide in-depth information about the synergistic effects of multiple features for the fault prediction model, and enable the fault prediction model to more accurately predict the sensitivity corresponding to the coordinate point.

[0082] Step 306: Input the feature vector into the fault prediction model to obtain the sensitivity corresponding to the coordinate point.

[0083] like Figure 4As shown, the fault prediction model in this embodiment is a prediction model based on a convolutional neural network. After the feature vector is reshaped into a tensor format suitable for one-dimensional convolution processing by a Reshape layer, it enters the feature extraction stage consisting of three progressive convolutional blocks: one-dimensional convolutional block 1, one-dimensional convolutional block 2, and one-dimensional convolutional block 3. The number of filters in the three progressive convolutional blocks increases in the order of 32, 64, and 128 to gradually extract higher-level feature representations. In each convolutional block, the one-dimensional convolutional layer Conv1D is responsible for capturing local feature patterns, the batch normalization layer BatchNorm ensures training stability, the max pooling layer MaxPool achieves dimensionality reduction and key feature preservation, and the random deactivation layer Dropout provides regularization to prevent overfitting. After the feature extraction stage is completed, the three-dimensional convolutional output is flattened into a 384-dimensional feature vector by the Flatten layer, and then enters the feature fusion stage consisting of two fully connected blocks, fully connected block 1 and fully connected block 2, to realize the transformation from local features to global decisions. Ultimately, the model generates continuous probability values ​​in the 0-1 range through an output layer equipped with a sigmoid activation function. These values ​​can be used directly for fault success rate prediction or converted into binary classification results by a threshold. For example, the threshold can be 0.5.

[0084] Step 308: Generate a sensitivity matrix based on the sensitivity of each coordinate point.

[0085] In this embodiment, feature vectors are obtained by performing feature enhancement processing on parameter vectors to reveal the complex coupling relationship between high-dimensional parameters. A fault prediction model based on convolutional neural networks is used to efficiently capture the local dependencies in the feature vectors, thereby constructing a high-precision fault prediction model verified by multi-dimensional indicators and improving the accuracy of the sensitivity of predicted coordinate points.

[0086] In one exemplary embodiment, based on Figure 2 The embodiment shown further includes: adjusting the target experimental conditions, returning to the step of inputting the target experimental conditions of each coordinate point in the spatial coordinate grid into a preset fault prediction model, and obtaining sensitivity heatmaps under multiple different target experimental conditions; performing point-by-point averaging on the multiple sensitivity heatmaps to obtain a comprehensive sensitivity map.

[0087] The target experimental conditions can be determined based on historical experimental conditions from historical electromagnetic fault injection experiments. Specifically, at least two power values ​​and at least two delay values ​​are selected from historical pulse power and delay values ​​from historical pulse delays, respectively, to form one or more target experimental conditions. For example, the target pulse power in one target experimental condition can be the 85th percentile of the historical pulse power distribution, and the target pulse delay can be the 85th percentile of the historical pulse delay distribution; the target pulse power in another target experimental condition can be the 95th percentile of the historical pulse power distribution, and the target pulse delay can be the 95th percentile of the historical pulse delay distribution.

[0088] In one possible implementation, the method may further include: normalizing the comprehensive sensitivity map to obtain a relative coordinate sensitivity map.

[0089] In some embodiments, the normalization process may be to linearly map the physical coordinate axis of the comprehensive sensitivity map to the relative coordinate space of the x-axis to generate a normalized relative coordinate sensitivity map.

[0090] In this embodiment, by averaging the sensitivity heatmaps under multiple different target experimental conditions point by point, a comprehensive sensitivity map is generated. This can integrate the influence of multiple experimental conditions, effectively weaken the accidental influence of a single experimental condition, and better reflect the regional vulnerability distribution of the target integrated circuit independent of specific parameters. The relative coordinate sensitivity map obtained by normalizing the comprehensive sensitivity map can eliminate the influence of the physical size of the target integrated circuit, which is convenient for cross-size comparative analysis, thereby further improving the comprehensiveness of the method for locating electromagnetically sensitive areas of integrated circuits.

[0091] In one exemplary embodiment, based on Figure 2 The embodiment shown further includes: acquiring an initial prediction model and raw data; standardizing the raw data to obtain experimental data; and training the initial prediction model using the experimental data to obtain a fault prediction model.

[0092] In some embodiments, the method further includes: standardizing the experimental data during the historical electromagnetic fault injection experiment to obtain standardized experimental data, and using the standardized experimental data to train the initial prediction model to obtain a fault prediction model.

[0093] In some embodiments, the method may further include: constructing a standardized model for subsequent standardization processing of new raw data.

[0094] For example, standardization can be based on Z-score standardization, Min-Max standardization, or Robust standardization.

[0095] In this embodiment, the process of training the initial prediction model may include: dividing the standardized experimental data into a training set and a validation set; using binary cross-entropy as the loss function and employing the Adam (Adaptive Moment Estimation) optimizer to train the initial prediction model on the training set. During training, an early stopping monitoring strategy is adopted to prevent overfitting based on the model's performance metrics on the validation set, such as validation loss, and to automatically terminate training when performance no longer improves, retaining the optimal model weights. The trained model undergoes a comprehensive performance evaluation on a reserved, untrained independent test set. The evaluation process can employ standard model evaluation metrics, specifically including: accuracy, precision, recall, F1 score, AUC (Area Under the Curve), and ROC curve (Receiver Operating Characteristic Curve).

[0096] In this embodiment, standardization of the original data eliminates dimensional differences. During model training, historical experimental conditions from the experimental data are used as input, and the resulting data is used as the output label. The model parameters are continuously adjusted using the backpropagation algorithm, enabling the model to accurately learn the mapping relationship between injection parameters and fault results. After training, the model is validated and optimized to ensure its prediction accuracy and generalization ability under different electromagnetic fault injection scenarios, thus providing reliable model support for subsequent fault prediction processing based on spatial coordinate grids. After obtaining the fault prediction model, near-zero-cost virtual injection analysis can be performed continuously at the software level, thus eliminating the need for expensive and time-consuming physical experiments. By adjusting the target experimental conditions, the corresponding sensitivity heatmap can be generated, greatly improving the analysis efficiency of the integrated circuit electromagnetic sensitive area location method.

[0097] In one exemplary embodiment, such as Figure 5 As shown, a method for locating electromagnetically sensitive areas of an integrated circuit is provided. The method is illustrated using an example of its application to locating electromagnetically sensitive areas of a microcontroller with a dual-core lockstep error handling mechanism. The method includes steps 501 to 512. Wherein:

[0098] Step 501: Obtain the initial prediction model and raw data.

[0099] Step 502: Standardize the raw data to obtain experimental data.

[0100] The experimental data includes historical experimental conditions and result data. Historical experimental conditions include historical pulse power and historical pulse delay. For example, a result data of Success=1 means "the dual-core lock-step error handling mechanism of the MCU was successfully triggered, for example, a specific error interrupt signal was generated." A total of 4858 experimental data points were collected, forming the dataset. Of these, 964 data points (approximately 20%) are the test set, and 3894 data points (approximately 80%) are the training set.

[0101] Step 503: Train the initial prediction model using experimental data to obtain the fault prediction model.

[0102] The fault prediction model in this embodiment performs well on the test set, as shown in Table 1 below. Specific ROC curves are shown in... Figure 6 .

[0103] Table 1 Training Results

[0104] accuracy Accuracy Recall rate F1 score AUC 0.89004 0.76942 0.95638 0.85277 0.94227

[0105] Step 504: Obtain the spatial coordinate grid and experimental conditions corresponding to the target integrated circuit.

[0106] The target experimental conditions include the target injection location coordinates, the target pulse power, and the target pulse delay.

[0107] Step 505: Generate parameter vectors corresponding to each coordinate point based on the target injection location coordinates, target pulse power, and target pulse delay.

[0108] Step 506: Perform feature enhancement processing on the parameter vector to obtain the feature vector.

[0109] The feature vector includes interactive features, polynomial features, and nonlinear features. Step 501 may further include: combining at least two features from the target injection position coordinates, target pulse power, and target pulse delay in the parameter vector, and performing multiplication operations in each combination to obtain multiple interactive features; performing exponentiation operations on the target injection position coordinates, target pulse power, and target pulse delay in the parameter vector to obtain multiple polynomial features; and transforming the target injection position coordinates, target pulse power, and target pulse delay using a preset nonlinear function to obtain multiple nonlinear features.

[0110] Step 507: Input the feature vector into the fault prediction model to obtain the sensitivity corresponding to the coordinate point.

[0111] Step 508: Generate a sensitivity matrix based on the sensitivity of each coordinate point.

[0112] Step 509: Generate a sensitivity heatmap corresponding to the target integrated circuit based on the sensitivity matrix.

[0113] Step 510: Adjust the target experimental conditions, return to step 505, and obtain sensitivity heatmaps under multiple different target experimental conditions.

[0114] Step 511: Perform point-by-point averaging on multiple sensitivity heatmaps to obtain a comprehensive sensitivity map.

[0115] Step 512: Normalize the comprehensive sensitivity map to obtain the relative coordinate sensitivity map.

[0116] For example, under target experimental condition 1: the target pulse power is at the 85th percentile of the historical pulse power distribution, and the target pulse delay is at the 85th percentile of the historical pulse delay distribution. The three-dimensional comprehensive sensitivity plot of the lockstep chip is shown in Figure 7(a), and the two-dimensional comprehensive sensitivity plot is shown in Figure 7(b). Under target experimental condition 2: the target pulse power is at the 85th percentile of the historical pulse power distribution, and the target pulse delay is at the 95th percentile of the historical pulse delay distribution. The three-dimensional comprehensive sensitivity plot of the lockstep chip is shown in Figure 8(a), and the two-dimensional comprehensive sensitivity plot is shown in Figure 8(b). ); Target experimental condition 3: The target pulse power is at the 95th percentile of the historical pulse power distribution, and the target pulse delay is at the 85th percentile of the historical pulse delay distribution. The three-dimensional plot of the overall sensitivity of the lockstep chip is shown in Figure 9(a), and the two-dimensional plot of the overall sensitivity is shown in Figure 9(b); Target experimental condition 4: The target pulse power is at the 95th percentile of the historical pulse power distribution, and the target pulse delay is at the 95th percentile of the historical pulse delay distribution. The three-dimensional plot of the overall sensitivity of the lockstep chip is shown in Figure 10(a), and the two-dimensional plot of the overall sensitivity is shown in Figure 10(b). The three-dimensional plot of the relative coordinate sensitivity map obtained under the above multiple target experimental conditions is shown in Figure 11(a); the two-dimensional plot of the relative coordinate sensitivity map is shown in Figure 11(b). Where sensitivity_score represents the sensitivity of the model output.

[0117] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0118] Based on the same inventive concept, this application also provides an integrated circuit electromagnetic sensitive area positioning device for implementing the aforementioned integrated circuit electromagnetic sensitive area positioning method. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more embodiments of the integrated circuit electromagnetic sensitive area positioning device provided below can be found in the limitations of the integrated circuit electromagnetic sensitive area positioning method described above, and will not be repeated here.

[0119] In one exemplary embodiment, such as Figure 12 As shown, an integrated circuit electromagnetic sensitive area positioning device is provided, including: a data acquisition module 1202, a matrix generation module 1204, and an image generation module 1206, wherein:

[0120] The data acquisition module 1202 is used to acquire the spatial coordinate grid and experimental conditions corresponding to the target integrated circuit.

[0121] The matrix generation module 1204 is used to input the target experimental conditions of each coordinate point in the spatial coordinate grid into the preset fault prediction model to obtain the sensitivity matrix corresponding to the spatial coordinate grid. The fault prediction model is trained based on the experimental data in the historical electromagnetic fault injection experiment process. The experimental data includes historical experimental conditions and result data. The historical experimental conditions include historical injection position coordinates, historical pulse power and historical pulse delay.

[0122] The image generation module 1206 is used to generate a sensitivity heatmap corresponding to the target integrated circuit based on the sensitivity matrix.

[0123] In one embodiment, the target experimental conditions include target injection location coordinates, target pulse power, and target pulse delay; the matrix generation module 1204 is further configured to generate parameter vectors corresponding to each coordinate point based on the target injection location coordinates, target pulse power, and target pulse delay; perform feature enhancement processing on the parameter vectors to obtain feature vectors; input the feature vectors into the fault prediction model to obtain the sensitivity corresponding to the coordinate points; and generate a sensitivity matrix based on the sensitivity corresponding to each coordinate point.

[0124] In one embodiment, the feature vector includes interactive features, polynomial features, and nonlinear features; the matrix generation module 1204 is further used to combine at least two features from the target injection position coordinates, target pulse power, and target pulse delay in the parameter vector, and perform multiplication operations in each combination to obtain multiple interactive features; to perform exponentiation operations on the target injection position coordinates, target pulse power, and target pulse delay in the parameter vector to obtain multiple polynomial features; and to perform transformation processing on the target injection position coordinates, target pulse power, and target pulse delay using a preset nonlinear function to obtain multiple nonlinear features.

[0125] In one embodiment, the matrix generation module 1204 is further used to adjust the target experimental conditions, return to the step of inputting the target experimental conditions of each coordinate point in the spatial coordinate grid into the preset fault prediction model, and obtain sensitivity heatmaps under multiple different target experimental conditions; perform point-by-point averaging on the multiple sensitivity heatmaps to obtain a comprehensive sensitivity map.

[0126] In one embodiment, the image generation module 1206 is further configured to normalize the comprehensive sensitivity map to obtain a relative coordinate sensitivity map.

[0127] In one embodiment, the device further includes: a model training module for acquiring an initial prediction model and raw data; standardizing the raw data to obtain experimental data; and training the initial prediction model using the experimental data to obtain a fault prediction model.

[0128] Each module in the aforementioned integrated circuit electromagnetic sensitive area positioning device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0129] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 13As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores spatial coordinate grids, target experimental conditions, and fault prediction models. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When executed by the processor, the computer program implements a method for locating electromagnetically sensitive areas of an integrated circuit.

[0130] Those skilled in the art will understand that Figure 13 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0131] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0132] In one exemplary embodiment, an integrated circuit electromagnetic sensitive area positioning system is also provided. The system includes a testing device and a computer device as described in the above embodiment. The testing device is connected to the computer device and is used to perform an electromagnetic fault injection experiment on the target integrated circuit, obtain experimental data, and send the experimental data to the computer device.

[0133] In one exemplary embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above-described method embodiments.

[0134] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0135] It should be noted that the data involved in this application (including but not limited to data used for analysis, data stored, data displayed, etc.) are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0136] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0137] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0138] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. An integrated circuit electromagnetic sensitive area positioning method, characterized by, The method comprises: acquiring a spatial coordinate grid corresponding to a target integrated circuit and target experimental conditions, the target experimental conditions comprising a target pulse power, a target pulse time delay, and a target injection position coordinate; generating a parameter vector corresponding to each coordinate point in the spatial coordinate grid according to the target injection position coordinate, the target pulse power, and the target pulse time delay; combining at least two features of the target injection position coordinate, the target pulse power, and the target pulse time delay in the parameter vector, and performing multiplication operation in each combination to obtain multiple interaction features; performing power operation on the target injection position coordinate, the target pulse power, and the target pulse time delay in the parameter vector respectively to obtain multiple polynomial features; performing transformation processing on the target injection position coordinate, the target pulse power, and the target pulse time delay respectively by using a preset nonlinear function to obtain multiple nonlinear features; inputting the interaction features, the polynomial features, and the nonlinear features into a fault prediction model to obtain a sensitivity corresponding to each coordinate point; obtaining a sensitivity matrix corresponding to the spatial coordinate grid according to the sensitivity corresponding to each coordinate point, the fault prediction model being obtained by training historical experimental data in an electromagnetic fault injection experiment, the experimental data comprising historical experimental conditions and result data, the historical experimental conditions comprising a historical pulse power, a historical pulse time delay, and a historical injection position coordinate; generating a sensitivity heat map corresponding to the target integrated circuit according to the sensitivity matrix.

2. The method of claim 1, wherein, The method further comprises: adjusting the target experimental conditions and returning to the step of inputting the target experimental conditions of each coordinate point in the spatial coordinate grid into a preset fault prediction model to obtain sensitivity heat maps under multiple different target experimental conditions; performing point-by-point average operation processing on the multiple sensitivity heat maps to obtain a comprehensive sensitivity map.

3. The method of claim 2, wherein, The method further comprises: performing normalization processing on the comprehensive sensitivity map to obtain a relative coordinate sensitivity map.

4. The method of claim 1, wherein, The method further comprises: acquiring an initial prediction model and original data; performing standardization processing on the original data to obtain the experimental data; training the initial prediction model by using the experimental data to obtain the fault prediction model.

5. The method of claim 3, wherein, The normalization processing on the comprehensive sensitivity map to obtain a relative coordinate sensitivity map comprises: linearly mapping a physical coordinate axis of the comprehensive sensitivity map to a relative coordinate space of an x-axis to generate the relative coordinate sensitivity map.

6. The method of claim 4, wherein, The method further comprises constructing a standardization model. The standardization processing on the original data to obtain the experimental data comprises: performing standardization processing on the original data by using the standardization model to obtain the experimental data.

7. An integrated circuit electromagnetic sensitive area positioning device, characterized by The device comprises: a data acquisition module configured to acquire a spatial coordinate grid corresponding to a target integrated circuit and target experimental conditions, the target experimental conditions comprising a target pulse power, a target pulse time delay, and a target injection position coordinate; The matrix generating module is configured to generate a parameter vector corresponding to each coordinate point in the spatial coordinate grid according to the target injection position coordinates, the target pulse power, and the target pulse time delay; combine at least two features among the target injection position coordinates, the target pulse power, and the target pulse time delay in the parameter vector, and perform multiplication operation in each combination to obtain a plurality of interaction features; perform power operation on the target injection position coordinates, the target pulse power, and the target pulse time delay in the parameter vector respectively to obtain a plurality of polynomial features; perform transformation processing on the target injection position coordinates, the target pulse power, and the target pulse time delay respectively by using a preset nonlinear function to obtain a plurality of nonlinear features; input the interaction features, the polynomial features, and the nonlinear features into a fault prediction model to obtain a sensitivity corresponding to each coordinate point; obtain a sensitivity matrix corresponding to the spatial coordinate grid according to the sensitivity corresponding to each coordinate point; the fault prediction model is obtained by training experimental data in a historical electromagnetic fault injection experiment; the experimental data includes historical experimental conditions and result data; the historical experimental conditions include historical pulse power, historical pulse time delay, and historical injection position coordinates; The image generating module is configured to generate a sensitivity heat map corresponding to the target integrated circuit according to the sensitivity matrix.

8. A computer device, comprising: The system comprises a memory and a processor, and the memory stores a computer program.

9. An integrated circuit electromagnetic sensitive area positioning system, characterized by, The system comprises a test device and the computer device of claim 8, the test device is connected with the computer device, and the test device is configured to perform an electromagnetic fault injection experiment on a target integrated circuit to obtain experimental data and send the experimental data to the computer device.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 6. The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 6.

Images