MATLAB: How to generate an Excel file from a Simulink Coverage object

coverageexcelsimulinkSimulink Check

I am using 'cvsim' and 'cvhtml' in the Command Window to generate Simulink coverage reports. Is there a way to generate an Excel file with a list of model elements missing coverage?

Best Answer

As of MATLAB R2020b, this is not possible, but you may be able to circumvent this limitation by writing a script.

Specifically, the '*info' functions (i.e. decisioninfo, conditioninfo, mcdcinfo, etc.) allow you to query coverage information from the 'cvdata' object and can be used to check if a specified block or model object is missing coverage.

A possible workaround would be to:

Get a list of all of the blocks in the model using find_system
Iterate over the list of blocks and for each, check the block's coverage using 'decisioninfo', 'conditioninfo', 'mcdcinfo', etc. (depending on what metrics you are interested in)
- In most cases, the *info function will return empty, meaning that this metric is not relevant for that block and this can be ignored.
- Otherwise, the function will return a two-element array showing the number of satisfied outcomes and the number total outcomes. If the former is less than the latter, then the block is missing coverage, and this can be logged as you see fit.

Related Solutions

MATLAB: Model Coverage Analysis for Logical Blocks

Decision Coverage in Simulink is defined as:

"Decision coverage analyzes elements that represent decision points in a model, such as a Switch block or Stateflow® states. For each item, decision coverage determines the percentage of the total number of simulation paths through the item that the simulation traversed."

Since the AND output is not driving a decision making block, there is no decision coverage to collect. By adding a switch that is being driven by the AND block, you can observe that the decision coverage is reported:

The AND block will need to have Condition coverage enabled to examine the level of coverage it has. Condition coverage verifies that all combinations of inputs are tested for a given Logical operator block.

MATLAB: How to extract code coverage metrics from a code coverage report

In order to achieve that you need to first create a test object.

>> testObj = cvtest(model);

Then you have to run code coverage on the model.

>> [cvdo,simOut] = cvsim(testObj);

Then you can get the coverage results by using the following functions:

>> blk_handle = get_param([model], 'Handle'); 
>> executionCoverage = executioninfo(cvdo, blk_handle);
>> conditionCoverage = conditioninfo(cvdo, blk_handle);
>> decisionCoverage  = decisioninfo(cvdo, blk_handle);
>> mcdcCoverage      = mcdcinfo(cvdo, blk_handle);

You can refer to the attached script which accomplishes the above workflow.

Please follow the documentation link below to automate additional coverage workflows:

https://www.mathworks.com/help/slcoverage/automate-coverage-workflows.html

Best Answer

Related Solutions

MATLAB: Model Coverage Analysis for Logical Blocks

MATLAB: How to extract code coverage metrics from a code coverage report

Related Question