Passing Strings Between Models

This example shows how to generate SimWB-compliant code for Simulink® models that exchange strings. The strings are converted to their ASCII equivalents and passed back and forth as uint8 arrays. Models exchange strings, for example, to simulate protocol handshakes and communicate commands.

Example Models
Generate SimWB-Compliant Code
Run Test Session in SimWB
Limitations

Example Models

This example uses two models, ccur_outstring and ccur_instring, which are both shipped with the MLToolkit in the $simwbroot$/examples folder. Open these models by typing ccur_outstring and ccur_instring at the MATLAB command prompt.

In ccur_outstring, the value of three source signals is encoded, using strings to specify the header and footer of each source signal. The encoded signal, which contains strings and numbers, is converted to its ASCII equivalent and output as an uint8 array. ccur_instring decodes this signal and decomposes it to display each signal's value.

The models show two ways of exchanging the encoded signal:

Using input/output blocks - You can use native input and output blocks to pass uint8 arrays. However, as these blocks support fixed-length arrays only, you can use this approach only for fixed-length strings. If you generate SimWB-compliant code using the MLToolkit GUI, the GUI will automatically handle these input/output blocks.
In this example, the port SWshareddata is used to pass the strings between the models.
Using toRTDBc and fromRTDBc blocks, with Treat as String option selected - Specify the array and array size as inputs/outputs of these blocks to pass variable-length uint8 arrays between models. Check the Treat as String box of the box. This ensures that the RTDB variables that are created use the STRI/STRO SimWB point type. If you generate SimWB-compliant code using the MLToolkit GUI, you must manually place and configure these blocks in your model before opening the GUI. After you generate the code, in SimWB, you must map the toRTDBc block to its corresponding fromRTDBc block using the RTDBItem-Copy I/O mapping. This mapping implicitly introduces a one-step lag in the string transmission.

Generate SimWB-Compliant Code

Because the two models share a signal, SWshareddata, before generating code using the MLToolkit GUI, you must modify the SimWB preferences. In the Simulink editor, select SimWB > Preferences. The SIMulation Workbench Preferences GUI opens. In the RTDB tab:

Uncheck the Hierarchical RTDB Variable Names check box - When you generate code for your model, the variables will not be prefixed by modelname.subsystem. Instead, the variable name will be the sink/source block name. So, ccur_outstring and ccur_instring will contain the SWshareddata variable in common.
Check the Append RTDB variables as AIO or DIO check box - When you append the RTDB for the second model, SimWB will register the common variables as AIOs (the example signals are all analog) instead of merely AI or AO. Note that, in a given model, a signal is either an input (analog or digital) or an output (analog or digital). When an RTDB registers a signal as an AIO, it means that this signal is used as an input in one model and an output in another model. This means that an RTDB variable is shared between models. An AIO or DIO signal does not mean that a signal is both input and output in the same model. That is, if a signal is input into a model, modified and then output, it does not qualify the signal as an AIO/DIO. In this case, configure the RTDB to use two variables. Each variable maps to a memory location. One memory location manages the input version of the signal, and another memory location manages the output version of the signal. This distinction between point types extends to bus and string signals as well.
Note that AIO/DIO/STRIO variables cannot be logged consistently.

Following the workflow described in Generate Code for Simulink Model Using MLToolkit, generate SimWB-Compliant Code for ccur_outstring. Load the code to an RTDB named ccur_string. For the Create RTDB Variables option, use the default settings, which will create RTDB variables for all source/sink blocks with the SW prefix.

Next, generate code for ccur_instring and load it to the same RTDB, ccur_string. Similar to ccur_outstring, when you create the RTDB variables, use the default settings for the Create RTDB Variables option. When you click Create and Upload RTDB, the RTDB Append/Overwrite dialog appears. Click Append/Update.

The signal.db file on the SimWB host machine (located in the relevant RTDB folder) shows that the SWshareddata variable has the AIO point type. This point type is derived based on how the signal is used in both models. However, the signal.db file on the local machine (located in the current MATLAB work directory) does not reflect the overall point type for this signal. Rather, it reflects the point type for the signals as relevant to the last model used by the MLToolkit GUI, ccur_instring.

Now that you have appended the RTDB and generated code for ccur_instring, you must regenerate the code for the first model, ccur_outstring. Regenerating the code ensures that it uses the latest signal point types.

Run Test Session in SimWB

Open the Simulation Workbench Control Center on your local machine. Load the ccur_string RTDB. To inspect its analog signals, in the editor menu, select RTDB & I/O > Analog Points.

Notice that the SWshareddata signal has the AIO point type. To inspect the string signals, in the editor menu, select RTDB & I/O > String Points.

Map mystringin to mystringout using the RTDBItem-Copy I/O Mapping.

First, add the RTDBItem-Copy an I/O device. In the editor menu, select RTDB & I/O > I/O Devices. From the Available Devices list, select RTDBItem-Copy. Click Add . Then, click Apply .

Next, in the editor menu, select RTDB & I/O > I/O Mappings. Open the RTDBItem-Copy-1 tab. Select mystringin. On the right-hand-side, check the mystringout check box to map the two variables to each other. Then, click Apply .

Create a test for the ccur_string RTDB.

Run a session of this test. Open the Real-Time Viewer by selecting Monitor > RT Viewer in the editor menu. Add all the analog variables to a chart and all the string variables to a table. You observe that the analog signals are decoded as expected and the strings are detected as expected.

Limitations

MLToolkit string support is available in MATLAB® R2010b and later releases only.
MLToolkit string support requires the use of C Mex S-functions. The source files for these S-functions are shipped with the MLToolkit. You must MEX these files for your specific MATLAB version and system platform. To make this process easy, use the createMex function. This function creates the MEX files for these S-functions and places them in the appropriate location. This function requires you to select the right compiler in MATLAB using mex -setup beforehand. To find the right compiler, refer to the MathWorks compiler selection table.