Automatic digital ic testing and performance evaluation method based on digital test platform

Abstract

The application discloses an automatic digital IC test and performance evaluation method based on a digital test platform, and relates to the technical field of IC testing.The steps of the method include the following steps: inserting a probe into a measured vehicle IC by using a digital test platform, simulating a real working condition of a vehicle, obtaining a benchmark performance, connecting a load throwing simulator, a sensor network and the measured IC, collecting vehicle voltage, temperature and CAN message load in real time, identifying a load throwing state, monitoring a vehicle state in real time, obtaining a vehicle state value, measuring IC execution delay, power consumption and error rate, calculating a performance state value, setting a complexity control algorithm, dynamically and adaptively adjusting the algorithm complexity step according to the performance sensitivity and the vehicle state, fitting a trend surface, identifying a critical complexity, and setting a complexity prediction model to output an optimal test complexity.The application solves the problems of low efficiency and high risk of fixed step or manual parameter adjustment caused by the great influence of algorithm complexity on IC performance during the vehicle IC test.

Description

Technical Field

[0001] This invention relates to the field of IC testing technology, specifically to an automated digital IC testing and performance evaluation method based on a digital testing platform. Background Technology

[0002] In vehicle electronic systems, digital ICs often undertake the operational tasks of critical functional modules, such as motor control, vehicle control, sensor data fusion, and communication message processing. Any performance abnormalities will directly affect the system's safety and stability. Automotive electronic systems place extremely high reliability requirements on microcontrollers, making early detection of design flaws crucial.

[0003] In real-world operation, digital ICs often face variable and complex operating environments, including sudden voltage fluctuations and instantaneous algorithm computation pressure. These factors have a multi-dimensional coupled impact on IC performance. Therefore, how to conduct high-precision, high-coverage dynamic performance testing and evaluation of digital ICs in environments close to real-world operating conditions has become a significant technical challenge for current digital chip design verification and quality control. Typical motor control closed loops require an update rate of 1-10kHz, and each computation must be completed within 100-1000μs; otherwise, the closed-loop stability will be severely compromised.

[0004] To this end, the present invention provides an automated digital IC testing and performance evaluation method based on a digital testing platform. Summary of the Invention

[0005] The purpose of this invention is to provide an automated digital IC testing and performance evaluation method based on a digital test platform to solve the existing problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an automated digital IC testing and performance evaluation method based on a digital test platform, comprising the following steps:

[0007] S1. Use a digital test platform to insert the test vehicle IC and simulate the real working conditions of the vehicle to obtain the baseline performance. Connect the load dump simulator, sensor network and IC under test.

[0008] S2. Real-time acquisition of vehicle voltage, temperature, and CAN message load, identification of load dump status, and real-time monitoring of vehicle status to obtain vehicle status values.

[0009] S3. Measure IC execution delay, power consumption, and error rate, and calculate performance status values;

[0010] S4. Set up a complexity control algorithm and dynamically and adaptively adjust the algorithm complexity step size according to performance sensitivity and vehicle status.

[0011] S5. Fit the trend surface, identify the critical complexity, and set the complexity prediction model to output the optimal test complexity.

[0012] A further improvement of this invention is that step S1 specifically includes loading firmware on the PXI, establishing a logical path between the vehicle IC under test and the digital platform, generating a standard voltage transient waveform through the load dump simulator interface, and simultaneously simulating CAN bus and embedded sensor network signals to generate vehicle operating condition waveforms; the embedded sensor network includes a high-speed voltage probe, a current detection chip, and a temperature sensor. The high-speed voltage probe is used to capture voltage transient spikes, the current detection chip is used to record the load current change rate, and the temperature sensor is used to record the IC temperature rise.

[0013] A further improvement of this invention is that the complexity control algorithm specifically includes the following steps:

[0014] S41. Establish vehicle test objects, including window control, air conditioning temperature control, light control and vehicle speed cruise control;

[0015] S42. Simultaneously run the test object and capture the vehicle status value Vs and performance status value psv in real time;

[0016] S43. Design independent PID parameters for each control task and generate PID parameter control sequences;

[0017] S44. Define the initial algorithm complexity as the weighted average of the number of operands and the amount of data in the PID parameter control sequence;

[0018] S45. Performance sensitivity is obtained by calculating the proportion of the performance state value to the change in the complexity of the previous algorithm. ;

[0019] S46. Set the initial complexity change step size. The algorithm calculates the first threshold Tvs1 and the second threshold Tvs2 for vehicle status, and updates the algorithm complexity step size in real time based on the real-time vehicle status value and performance sensitivity.

[0020] S47. Establish a weight update model to adjust the performance sensitivity and vehicle state value weights in real time.

[0021] A further improvement of the present invention is that step S46 specifically includes:

[0022] When the vehicle state value is less than or equal to the first threshold of the vehicle state, it is judged to be in a safe state. At this time, the algorithm complexity step size is... Maintaining linear decay, the calculation formula is expressed as: ,in, Indicates the minimum allowable step size. Indicates the maximum allowed step size;

[0023] When the vehicle status value is greater than the first threshold or less than the second threshold, it is judged as a warning state. At this point, the algorithm complexity step size... The updated formula is ,in, and This indicates the set weight. Indicates the first This adjustment;

[0024] When the vehicle status value is greater than or equal to the second threshold of the vehicle status, it is judged to be in an overload state, and the test is stopped at this time.

[0025] The first threshold for vehicle status is less than the second threshold for vehicle status.

[0026] A further improvement of this invention is that the weight update model specifically includes:

[0027] The vehicle control scenario is mapped onto a reinforcement learning framework, including vehicle state values, performance state values, and the current time t, to obtain the current time control state vector. and will Data normalization This indicates the trend of performance status changes. This indicates the trend of changes in vehicle status. and This represents the weights learned in the previous moment;

[0028] Computational performance is close to the reward. , This represents the performance baseline value, used to penalize the absolute distance from the target deviation; and sets the smoothness penalty. Receive total reward ;

[0029] Will As input, the SAC algorithm is iterated for 10 iterations. 5 Step, output the mean and variance parameters to obtain the sampling action. Converted into weight increment , , Indicates the adjustment rate; training sampling action The two Critic networks using SAC calculate and update the Q-value of each sampling action based on the total reward. The sampling action network optimizes the gradient direction provided by the Critic and outputs new weights.

[0030] A further improvement of the present invention is that step S5 specifically includes:

[0031] S51. The fitting performance of quadratic functions varies non-linearly with complexity;

[0032] S52. Take the derivative of the fitted curve to find the algorithm complexity corresponding to the minimum performance state, and obtain the first critical complexity. And find the algorithm complexity corresponding to the standard performance value as the second critical complexity; ;

[0033] S53. Using historical algorithm complexity and vehicle state values ​​as input features, and performance state values ​​as the target variable, a support vector regression model is used to predict performance state values ​​under different algorithm complexities in real time. ;

[0034] S54. If the actual performance state value is less than the predicted performance state value three times consecutively, the first critical complexity update is triggered. ;

[0035] S55, Output The algorithm complexity corresponding to the maximum value is sent to the terminal as the optimal complexity of the vehicle IC.

[0036] A further improvement of this invention is that the first critical complexity is expressed by the formula If obtained, then ,in, Indicate the complexity of the historical algorithm; if a If 0, then it represents the algorithm complexity corresponding to the minimum performance state. If the result is 0, then refitting is required.

[0037] A further improvement of this invention is that the specific step S2 includes capturing the bus voltage, calculating the smoothed voltage MA and transient voltage gradient ΔV in real time, and then weighting and summing the normalized smoothed voltage and transient voltage gradient to obtain the voltage risk. The CAN analyzer reports the CAN bus occupancy rate every 10ms. Normalization yields bus occupancy risk The temperature sensor reports the real-time temperature Tamb every 100ms, and the temperature risk is obtained by normalization. ; Obtain vehicle status value ,in , and The weights are represented; and Vs is calculated every 10ms and written to a shared variable.

[0038] A further improvement of this invention is that the performance status value is obtained by setting performance monitoring targets, including execution latency, dynamic power consumption, and CAN frame error rate; firstly, the execution latency is obtained by measuring the time from "task triggering" to "result output" using a built-in 1ns resolution timing module. Secondly, the average power of voltage and current at N=1000 points is collected to obtain the dynamic power consumption. The CAN frame error rate is obtained by collecting the percentage of erroneous frames in the total number of transmitted frames; the performance status value is obtained by weighted summation of the normalized execution delay, dynamic power consumption, and CAN frame error rate.

[0039] Compared with the prior art, the beneficial effects of the present invention are:

[0040] 1. This invention first designs a dynamic step size update strategy based on performance sensitivity and vehicle state values, which allows for rapid adjustment of complexity in the safe zone, exponential decay of the step size in the warning zone, and cessation of testing in the overload zone, thus constructing a complete "safety first, fine control" mechanism.

[0041] 2. By mapping complexity control to the Soft Actor-Critic reinforcement learning framework, online intelligent optimization of weights is achieved, enabling step size adjustment to learn and adapt to performance and vehicle trends. Attached Figure Description

[0042] Figure 1 This is a flowchart of the automated digital IC testing and performance evaluation method based on a digital testing platform according to the present invention;

[0043] Figure 2 This is a flowchart of the complexity control algorithm of the present invention;

[0044] Figure 3 This is a flowchart of the complexity prediction model of the present invention. Detailed Implementation

[0045] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0046] The term "and / or" simply describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0047] Example 1

[0048] Figure 1 The flowchart of the automated digital IC testing and performance evaluation method based on a digital test platform disclosed in this embodiment is shown. The steps are as follows:

[0049] S1. A digital test platform is used to test the vehicle-mounted IC under test (V2T) and simulate real vehicle operating conditions to obtain baseline performance. A load dump simulator, sensor network, and V2T are connected. Specifically, firmware is loaded onto the PXI, a logical path is established between the V2T and the digital platform, a standard voltage transient waveform is generated through the load dump simulator interface, and simultaneously, CAN bus and embedded sensor network signals are simulated to generate vehicle operating condition waveforms. The embedded sensor network includes a high-speed voltage probe, a current detection chip, and a temperature sensor. The high-speed voltage probe captures voltage transient spikes, the current detection chip records the load current change rate, and the temperature sensor records the IC temperature rise. This process is used to reproduce vehicle conditions.

[0050] The high-voltage probe uses a high-spring-force probe to directly press against the vehicle-mounted IC pins; the load dump simulator interface connects a programmable pulse source between the 12V main bus and the DUT (Device Under Test); the programmable pulse source uses a digital signal controller (DSC) to drive the switching circuit, generating a load dump waveform that meets the requirements. , , The exponential decay simulates the natural recovery process of the load dump voltage. It is a time constant. .

[0051] S2. Real-time acquisition of vehicle voltage, temperature, and CAN message load; identification of load dumping states; and real-time monitoring of vehicle status to obtain vehicle status values. Specific steps include capturing bus voltage, real-time calculation of the smoothed voltage MA and transient voltage gradient ΔV, and weighted summation of the normalized smoothed voltage and transient voltage gradient to obtain voltage risk. The CAN analyzer reports the CAN bus occupancy rate every 10ms. Normalization yields bus occupancy risk The temperature sensor reports the real-time temperature Tamb every 100ms, and the temperature risk is obtained by normalization. ; Obtain vehicle status value ,in , and The weights are represented; and Vs is calculated every 10ms and written to a shared variable; where , This represents the voltage within the window, W = 10 ms; .

[0052] S3. Measure the IC execution delay, power consumption, and error rate to calculate performance status values. These values ​​are obtained by setting performance monitoring targets, including execution delay, dynamic power consumption, and CAN frame error rate. First, the execution delay is obtained by measuring the time from "task trigger" to "result output" using the built-in 1ns resolution timing module. Secondly, the average power of voltage and current at N=1000 points is collected to obtain the dynamic power consumption. , The CAN frame error rate is obtained by collecting the percentage of erroneous frames in the total number of transmitted frames; the performance status value is obtained by weighted summation of the normalized execution delay, dynamic power consumption, and CAN frame error rate.

[0053] S4. Set up a complexity control algorithm and dynamically and adaptively adjust the algorithm complexity step size based on performance sensitivity and vehicle status.

[0054] S5. Fit the trend surface, identify the critical complexity, and set the complexity prediction model to output the optimal test complexity.

[0055] Example 2

[0056] Based on the inventive concept of Embodiment 1, this embodiment provides specific implementation steps of the complexity control algorithm in an automated digital IC testing and performance evaluation method based on a digital testing platform. Figure 2 The flowchart of the complexity control algorithm of this invention is shown, and the specific steps include:

[0057] S41. To ensure the safety and efficiency of vehicle ICs, and to adaptively select the algorithm complexity according to the vehicle's own state, test objects are established, including window control, air conditioning temperature control, light control, and vehicle speed cruise.

[0058] S42. Simultaneously run the test object and capture the vehicle status value Vs and performance status value psv in real time;

[0059] S43. Design independent PID parameters for each control task and generate PID parameter control sequences;

[0060] S44. The PID control algorithm itself is simple, but its complexity mainly stems from the characteristics of the controlled object and the real-time requirements. A high sampling frequency means the controller needs to complete calculations in a shorter time, demanding higher processing speeds and potentially requiring algorithm structure optimization. Therefore, the initial algorithm complexity is defined as the weighted average of the number of operands and the amount of data in the PID parameter control sequence. For example, the operand for window control is the number of operations, and the amount of data is the standard value of window movement; the operand for air conditioning temperature control is the number of operations, and the amount of data is the standard value of window temperature adjustment; the operand for headlight control is the number of operations, and the amount of data is the standard value of headlight brightness change; the operand for cruise control is the number of gear changes, and the amount of data is the standard value of speed change.

[0061] S45. Performance sensitivity is obtained by calculating the proportion of the performance state value to the change in the complexity of the previous algorithm. ;

[0062] S46. Continuing to test after the load has been removed will damage the chip; if the algorithm complexity is applied in large steps under high load, it will still damage the chip and the optimal complexity cannot be found.

[0063] Therefore, an initial step size for the complexity change is set. The algorithm calculates the first threshold Tvs1 and the second threshold Tvs2 for vehicle status, and updates the algorithm complexity step size in real time based on the real-time vehicle status value and performance sensitivity.

[0064] When the vehicle state value is less than or equal to the first threshold, it is considered to be in a safe state. At this point, the vehicle state is healthy, and the vehicle state adjustment step size can be disregarded. Therefore, the algorithm complexity step size is... Maintain linear decay, ,in, Indicates the minimum allowable step size. Indicates the maximum allowed step size;

[0065] When the vehicle status value is greater than the first threshold and less than the second threshold, it is judged as a warning state. At this point, the algorithm complexity step size is... The updated formula is ,in, and This indicates the set weight. Indicates the first This adjustment;

[0066] When the vehicle status value is greater than or equal to the second threshold of the vehicle status, it is judged to be in an overload state, and the test is stopped at this time.

[0067] The first threshold for vehicle status is less than the second threshold for vehicle status.

[0068] This embodiment achieves smooth adjustment through an exponential function under normal conditions and rapid complexity reduction through a step transition under emergency conditions; it also accelerates the response while limiting the upper limit to prevent oscillations.

[0069] S47. Establish a weight update model to adjust the weights of performance sensitivity and vehicle state values ​​in real time; the weight update model specifically includes:

[0070] The vehicle control scenario is mapped onto a reinforcement learning framework, including vehicle state values, performance state values, and the current time t, to obtain the current time control state vector. and will Data normalization This indicates the trend of performance status changes. This indicates the trend of changes in vehicle status. and This represents the weights learned in the previous moment;

[0071] Computational performance is close to the reward. , This represents the performance baseline value, used to penalize the absolute distance from the target deviation; and sets the smoothness penalty. Receive total reward ;

[0072] Will As input, the SAC algorithm is iterated for 10 iterations. 5 Step, output the mean and variance parameters to obtain the sampling action. Converted into weight increment , , Indicates the adjustment rate; training sampling action The two Critic networks using SAC calculate and update the Q-value of each sampling action based on the total reward. The sampling action network optimizes the gradient direction provided by the Critic and outputs new weights.

[0073] This embodiment increases the weight of vehicle state values ​​when the trend of increasing vehicle state values ​​is greater, and increases the weight of performance state sensitivity when the trend of increasing performance state values ​​is greater. This increases the flexibility of weight changes and takes into account the balance of step size changes in calculation.

[0074] Example 3

[0075] Based on the inventive concepts of Embodiments 1 and 2, this embodiment provides specific implementation steps for the complexity prediction model in an automated digital IC testing and performance evaluation method based on a digital testing platform. Figure 3 The flowchart of the complexity prediction model of this invention is shown, and the specific steps include:

[0076] S51. The fitting performance of quadratic functions varies non-linearly with complexity;

[0077] S52. Take the derivative of the fitted curve to find the algorithm complexity corresponding to the minimum performance state, and obtain the first critical complexity. This represents the minimum acceptable complexity threshold, where further increases in complexity are not advisable; and the algorithm complexity corresponding to the standard performance value is identified as the second critical complexity. The first critical complexity is expressed by the formula. If obtained, then ,in, Indicate the complexity of the historical algorithm; if a If 0, then it represents the algorithm complexity corresponding to the minimum performance state. If the result is 0, then refitting is required;

[0078] S53. Using historical algorithm complexity and vehicle state values ​​as input features, and performance state values ​​as the target variable, a support vector regression model is used to predict performance state values ​​under different algorithm complexities in real time. ;

[0079] S54. If the actual performance state value is less than the predicted performance state value three times consecutively, the first critical complexity update is triggered. ;

[0080] S55, Output The algorithm complexity corresponding to the maximum value is sent to the terminal as the optimal complexity of the vehicle IC.

[0081] The threshold and weight settings can be set by default according to the present invention, or they can be set by the operator.

[0082] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0083] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0084] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0085] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0086] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. An automated digital IC testing and performance evaluation method based on a digital test platform, characterized by: Includes the following steps: S1. Use a digital test platform to insert the test vehicle IC and simulate the real working conditions of the vehicle to obtain the baseline performance. Connect the load dump simulator, sensor network and IC under test. S2. Real-time acquisition of vehicle voltage, temperature, and CAN message load, identification of load dump status, and real-time monitoring of vehicle status to obtain vehicle status values. S3. Measure IC execution delay, power consumption, and error rate, and calculate performance status values; S4. Set up a complexity control algorithm and dynamically and adaptively adjust the algorithm complexity step size according to performance sensitivity and vehicle status. The specific steps of the complexity control algorithm include: S41. Establish vehicle test objects, including window control, air conditioning temperature control, light control and vehicle speed cruise control; S42. Simultaneously run the test object and capture the vehicle status value Vs and performance status value psv in real time; S43. Design independent PID parameters for each control task and generate PID parameter control sequences; S44. Define the initial algorithm complexity as the weighted average of the number of operands and the amount of data in the PID parameter control sequence; S45. Performance sensitivity is obtained by calculating the proportion of the performance state value to the change in the complexity of the previous algorithm. ; S46. Set the initial complexity change step size. The algorithm calculates the first threshold Tvs1 and the second threshold Tvs2 for vehicle status, and updates the algorithm complexity step size in real time based on the real-time vehicle status value and performance sensitivity. S47. Establish a weight update model to adjust the weights of performance sensitivity and vehicle state value in real time. S5. Fit the trend surface, identify the critical complexity, and set the complexity prediction model to output the optimal test complexity.

2. The automated digital IC testing and performance evaluation method based on a digital test platform according to claim 1, characterized in that: Step S1 specifically includes loading firmware on the PXI, establishing a logical path between the vehicle IC under test and the digital platform, generating a standard voltage transient waveform through the load dump simulator interface, and simultaneously simulating CAN bus and embedded sensor network signals to generate vehicle operating condition waveforms. The embedded sensor network includes a high-speed voltage probe, a current detection chip, and a temperature sensor. The high-speed voltage probe is used to capture voltage transient spikes, the current detection chip is used to record the load current change rate, and the temperature sensor is used to record the IC temperature rise.

3. The automated digital IC testing and performance evaluation method based on a digital test platform according to claim 1, characterized in that: Step S46 includes the following specific steps: When the vehicle state value is less than or equal to the first threshold, it is considered to be in a safe state. At this point, the algorithm's complexity step size... Maintaining linear decay, the calculation formula is expressed as: ,in, Indicates the minimum allowable step size. Indicates the maximum allowed step size; When the vehicle status value is greater than the first threshold and less than the second threshold, it is judged as a warning state. At this point, the algorithm complexity step size is... The updated formula is ,in, and This indicates the set weight. Indicates the first This adjustment; When the vehicle status value is greater than or equal to the second threshold of the vehicle status, it is judged to be in an overload state, and the test is stopped at this time. The first threshold for vehicle status is less than the second threshold for vehicle status.

4. The automated digital IC testing and performance evaluation method based on a digital test platform according to claim 1, characterized in that: The weight update model specifically includes: The vehicle control scenario is mapped onto a reinforcement learning framework, including vehicle state values, performance state values, and the current time t, to obtain the current time control state vector. and will Data normalization This indicates the trend of performance status changes. This indicates the trend of changes in vehicle status. and This represents the weights learned in the previous moment; Computational performance is close to the reward. , This represents the performance baseline value, used to penalize the absolute distance from the target deviation; and sets the smoothness penalty. Receive total reward ; Will As input, the SAC algorithm is iterated for 10 iterations. 5 Step, output the mean and variance parameters to obtain the sampling action. Converted into weight increment , , Indicates the adjustment rate; training sampling action The two Critic networks using SAC calculate and update the Q-value of each sampling action based on the total reward. The sampling action network optimizes the gradient direction provided by the Critic and outputs new weights.

5. The automated digital IC testing and performance evaluation method based on a digital test platform according to claim 1, characterized in that: The specific steps of S5 include: S51. The fitting performance of quadratic functions varies non-linearly with complexity; S52. Take the derivative of the fitted curve to find the algorithm complexity corresponding to the minimum performance state, and obtain the first critical complexity. And find the algorithm complexity corresponding to the standard performance value as the second critical complexity; ; S53. Using historical algorithm complexity and vehicle state values ​​as input features, and performance state values ​​as the target variable, a support vector regression model is used to predict performance state values ​​under different algorithm complexities in real time. ; S54. If the actual performance state value is less than the predicted performance state value three times consecutively, the first critical complexity update is triggered. ; S55, Output The algorithm complexity corresponding to the maximum value is sent to the terminal as the optimal complexity of the vehicle IC.

6. The automated digital IC testing and performance evaluation method based on a digital test platform according to claim 5, characterized in that: The first critical complexity is expressed by the formula If obtained, then ,in, Indicate the complexity of the historical algorithm; if a If 0, then it represents the algorithm complexity corresponding to the minimum performance state. If the result is 0, then refitting is required.

7. The automated digital IC testing and performance evaluation method based on a digital test platform according to claim 1, characterized in that: The specific steps of S2 include capturing the bus voltage, calculating the smoothed voltage MA and transient voltage gradient ΔV in real time, and then normalizing the smoothed voltage and transient voltage gradient before weighted summation to obtain the voltage risk. The CAN analyzer reports the CAN bus occupancy rate every 10ms. Normalization yields bus occupancy risk The temperature sensor reports the real-time temperature Tamb every 100ms, and the temperature risk is obtained by normalization. ; Obtain vehicle status value ,in , and The weights are represented; and Vs is calculated every 10ms and written to a shared variable.

8. The automated digital IC testing and performance evaluation method based on a digital test platform according to claim 1, characterized in that: The performance status values ​​are obtained by setting performance monitoring targets, including execution latency, dynamic power consumption, and CAN frame error rate; firstly, the execution latency is obtained by measuring the time from task triggering to result output using the built-in 1ns resolution timing module. Secondly, the average power of voltage and current at N=1000 points is collected to obtain the dynamic power consumption. The CAN frame error rate is obtained by collecting the percentage of erroneous frames in the total number of transmitted frames; the performance status value is obtained by weighted summation of the normalized execution delay, dynamic power consumption, and CAN frame error rate.

Images