Automated Generation Of Code From Simulink To An Event-Based Simulation Framework

Size: px

Start display at page:

Download "Automated Generation Of Code From Simulink To An Event-Based Simulation Framework"

Tracy Williams
5 years ago
Views:

1 Automated Generation Of Code From Simulink To An Event-Based Simulation Framework Jonathan Maram Isela Villanueva Millennium Engineering And Integration 16 March 2010 AKC SE Plan

2 Simulation Environments Used By MEI Simulink Dynamic System Modeling/Simulations Graphical Block-Diagramming Tool Tightly Integrated With Matlab Command Library Means To Generate Time-Driven Code Event-based Framework (EBF) Discrete Event Messaging Models created with compiled code Componentized Simulation Layers Middleware, Controller Layers User Components; Simulation Engine & Model Layers Benefit Derived From Porting Simulink-Developed Models Into EBF Simulations 2

3 Required Information Flow From Simulink To EBF During Code Generation Simulink Model Logic Model Traits Model1 Out1 In2 Model Logic Model Traits Model2 EBF Code Elements Model1 Constructor Model1 Initializer Model1 Event1 Constructor Model1 Self-Schedule Event Handler Model2 Constructor Model2 Initializer Model1 Event1 Handler 3

4 Time-consuming Manual Approach Initially Used Functional model code generated from Simulink Other components hand-written Functional Model (Code Generated from Simulink or other sources) Implements Methods in the Model API Input User XML Output Graphic File Database Model API Model Base Class Inline Functionality API Interface Code Process Events, Call Functional Model Publish and Subscribe to Attributes, Output Streams Implements Methods in the Inline Functionality API Simulation Engine Controller 4

5 Tasks For Automation Extract Information About Targeted Models And Their I/O Store Information In A.MAT file Generate Functional Model Code From Simulink Using Standard Templates Generate Additional Interface Code Using Custom Code Templates Model classes, including Event Handling Methods Attribute classes Event classes Generate Makefile Or Visual Studio Projects/Solution Compile & Build EBF code Generate Prototype Run-Time Configuration Files Incorporate Simulink Tunable Parameters 5

6 Automated EBF Code Generation Implemented Reusable Code Templates Matlab Scripts EBF API Model Development Develop ETE Simulink Model Run Simulink Simulations Configure For CG Extracted Model Data Tunable Parameters Generate Custom Code Templates Generate Code Baseline Config Files (Customizations) Automated Code Generation VS Solution VS Projects C++ files Compare Results Run EBF Simulations Model DLLs Simitar Executable Build EBF Simulation 6

7 Testing Performed To Verify Fidelity Of Generated Code 4 Synthetic Test Cases varied I/O schemes varied feedback schemes multiple rates 4 Simulink Models Developed In-House or By MEI Partners Representation of Sensors, Association, State Estimation, Control, Actuators, and Flight Dynamics Output data from Simulink & EBF Simulations Compared 7

8 Example Scenario Targeted For Code Generation 8

9 Example Parameter Compared: Pointing Error 9

10 Pointing Error Time Plots Compared 10

11 Pointing Errors Compare Well 11

12 Summary Customized scripts and templates used to generate EBF scenario Inputs/Outputs Characterized and Adapted Model Traits Captured And Incorporated Templates Structured With EBF-Specific Code Model, Event, Attribute Classes Model Initialization and Event Handling Methods Communication, Event Scheduling, Data Handling Code Generation Interface Tested and Validated With Several Scenarios 12

Modeling HDL components for FPGAs in control applications

Modeling HDL components for FPGAs in control applications Mark Corless, Principal Application Engineer, Novi MI 2014 The MathWorks, Inc. 1 Position sensing High resolution voltage modulation Critical diagnostics