Processor-in-the-Loop Verification of MATLAB Functions Using STMicroelectronics STM32 Processors

Prerequisites

Required Hardware

Create a New Folder and Copy Relevant Files

In this example, you are going to verify a MATLAB function that adds two inputs and returns output. Create a new folder and copy relevant files which required to verify on ARM Cortex-M4 hardware board..

Navigate to new folder and create a function file, simple_addition.m, containing a function that adds two inputs and returns the result.

To see the contents of this function:

type simple_addition

The %#codegen directive in this function indicates that the MATLAB code is intended for code generation.

Generate PIL MEX Function from Command Line

The following steps provide commands to generate code and a library for a MATLAB function to verify on STM32 Nucleo-F401RE board.

Step 1: Create a coder.EmbeddedCodeConfig object to generate code and create a library for MATLAB function simple_addition.

config = coder.config('lib','ecoder',true);

Step 2: Configure the object for PIL.

config.VerificationMode = 'PIL';

Step 3: Specify the hardware on which the generated code to be verified.

config.Hardware = coder.hardware('STM32F4xx Based');

Step 4: Configure the stm32_MATLAB_PIL.ioc file. For more information, see Getting Started with STMicroelectronics STM32 Processor Based Boards

Step 5: Choose the PIL communication interface.

To choose 'ST-LINK' as PIL communication interface, below is the step:

config.Hardware.PILInterface = 'ST-LINK';

To choose 'Serial (USART2)' as PIL communication interface, below are the steps.

config.Hardware.PILInterface = 'Serial (USART2)';
config.Hardware.PILCOMPort = 'COM28';

Provide the PILCOMPort of your machine. The com port provided is just an example and might vary depending on individual machine. To see the list of available COM ports on your computer, select Start > Control Panel > Device Manager > Ports (COM & LPT)

Step 6: Limit the stack size to reasonable size, for example 512 bytes, as default size is much larger than the memory available on the hardware.

config.StackUsageMax = 512;

Step 7: You can enable verbose build to view the build log on command line.

config.Verbose = 1;

Step 8: Generate library code for the simple_addition MATLAB function and the PIL interface.

inp = single(zeros(1,30)); codegen('-config', config, '-args',
{inp,inp}, 'simple_addition');

In above commands, inp declares the data type and size for input arguments to MATLAB function 'simple_addition'. The codegen command generates code into following folders

Also, this step creates simple_addition_pil PIL MEX function in the current folder. This allows you to test the MATLAB code and the PIL MEX function and compare the results between both.

Note: You can also use the build_pil_mex function to generate the PIL mex function for any MATLAB function. The required input arguments are:

1. fcName - Function name for which code is to be generated

2. Inputs - Inputs declaring the data type and size for input arguments as a cell array.

3. hwName - Name of the Hardware board

Eg: To generate code for simple_addition function for STM32 Nucleo-F401RE hardware :

inp = single(zeros(1,30));
build_pil_mex('simple_addition',{inp,inp},'STM32F4xx Based');

You can also specify the COM port and baud rate by passing them as name and value pairs. Default PIL COM port is COM28 and PIL baud rate is 115200.

build_pil_mex('simple_addition',{inp,inp},'STM32F4xx Based', 'PILCOMPort','COM6', 'PILBaudRate','9600');

Run the PIL MEX Function

Run the PIL MEX function to compare its behavior to that of the original MATLAB function and to check for run-time errors.

u1 = single(rand(1,30)); u2 = single(rand(1,30)); y =
simple_addition_pil(u1,u2);

Terminate PIL execution with the following command.

clear simple_addition_pil;

Verify Generated Code

To verify the numerical accuracy of the generated code, compare MATLAB results with those of the PIL MEX function:

norm(y - simple_addition(u1,u2))

Profile Generated Code

To enable code execution profiling, set CodeExecutionProfiling of the coder.EmbeddedCodeConfig coder configuration object to true before generating code for the MATLAB function in Step 7 of Generate PIL MEX Function from Command Line.

config.CodeExecutionProfiling = true;

When profiling is enabled, the generated code is instrumented with timing information. The profiling results are transferred to MATLAB when the PIL MEX function is cleared from memory.

To accumulate profiling results, run simple_addition function 100 times in a loop:

for k=1:100, y = simple_addition_pil(u1,u2); end

The profiling results are available after clearing the PIL MEX function:

clear simple_addition_pil

Bring up the profiling report:

ProfileResultsWithoutCRL = getCoderExecutionProfile('simple_addition');
report(ProfileResultsWithoutCRL)

Use ARM Cortex-M Code Replacement Library (CRL)

To take advantage of optimized CRL for ARM Cortex-M processors, assign CodeReplacementLibrary library 'ARM Cortex-M (CMSIS)' and re-build the PIL MEX function following Step 7 of Generate PIL MEX Function from Command Line:

config.CodeReplacementLibrary = 'ARM Cortex-M (CMSIS)';

Note: If using build_pil_mex function, you can enable CRL replacements by specifying enableCRL as true .

build_pil_mex('simple_addition',{inp,inp},'STM32F4xx Based', 'PILCOMPort','COM6', 'PILBaudRate','9600','enableCRL', true) ;

Run the simple_addition function 100 times in a loop to accumulate profiling results:

for k=1:100, y = simple_addition_pil(u1,u2); end

The profiling results are available after clearing the PIL MEX function:

clear simple_addition_pil

Bring up the profiling report:

ProfileResultsWithCRL = getCoderExecutionProfile('simple_addition');
report(ProfileResultsWithCRL)

Compare the profiling results to those obtained in Profile Generated Code.

More About