Model-based tools solve fixed-point design

Model-based design tools from The MathWorks have enabled Motorola to accurately and efficiently create embedded software for fixed-point controllers, solving major design challenges.

Model-based design tools from The MathWorks have enabled Motorola to accurately and efficiently create embedded software for fixed-point controllers, solving major design challenges.

Using Simulink, Stateflow and Real-Time Workshop Embedded Coder, Motorola designed and automatically generated production code for its controllers, resulting in significant reductions in both development time and costs.

Motorola is the latest company in a list of customers that are using these tools to generate automatic production code, including Jaguar, BAE Systems and Honeywell.

Motorola's Automotive Group used The MathWorks' software to develop an electronic control unit that manages the battery of an electric vehicle.

The automatically generated software runs on Motorola's 16bit, fixed-point M68HC12 at 16MHz, with 128Kbit of Flash memory and 8Kbit of RAM.

Salam Zeiden of Motorola said, "The MathWorks model-based design and production code generation software were heavily used to develop and optimise the battery management controller software.

We most likely would not have been able to meet our project's delivery dates if we had used traditional methods of paper designs and hand coding instead of using Simulink, Stateflow, and Real-Time Workshop Embedded Coder." The MathWorks' model-based design software enables embedded systems developers to design software by creating and refining a model of the system.

They can simulate, test, and analyse the model and automatically generate target-ready floating-point or fixed-point C code for the system at any point in the development process.

The process is faster and less error prone than hand coding the embedded control software.

The model-based design software includes fully integrated support for both fixed-point and floating-point data types.

This enables engineers to design algorithms, simulate their behaviour, assess tradeoffs between floating-point and fixed-point designs, and automatically generate extremely efficient implementations for either floating-point or fixed-point controllers.

Embedded target products also make it easy to compile and execute code on the embedded processors by directly supporting specific microcontroller units and DSP chips for rapid prototyping, production deployment of embedded applications, and validation and performance analysis.

Additional production code generation capabilities make it easier to develop software for large systems, while autoscale simulation and data type propagation capabilities can quickly optimise fixed-point scaling values based on numerical stress tests.

As a result, tremendous time and effort savings are achieved.

Back to top