The Potential of Model-Based Deveopment (MBD) and Key Points for Its Introduction

Recently, the keyword 'CASE'—which stands for Connected, Autonomous, Shared & Services, and Electric—has been gaining prominence in the automotive industry. Particularly, the development of automotive control systems is reaching a significant turning point in realizing these technologies. Leading this revolution is Model-Based Development (MBD).

Compared to the traditional approach of 'creating actual parts, assembling them, and testing the prototype,' Model-Based Development (MBD) uses 'virtual models' to represent 'reality.'
Specifically, by using tools such as MATLAB/Simulink to repeatedly simulate the behavior of prototypes created in a virtual space, MBD achieves efficient development.

What benefits can be expected from 'modeling' in control system development?

In traditional methods, various departments work on the design, implementation, and testing of different parts of a system prototype using their own personnel and budget. Afterward, these parts are assembled into a prototype, which is then evaluated through integration tests and system tests. Each department takes the evaluation results back and makes improvements for the next prototype. Achieving a fully functional system on the first attempt is extremely challenging with this method, and it typically requires multiple iterations to improve the prototype’s performance.

Modeling enables the following outcomes:

• Centralized management of the entire process allows for efficient information sharing among all developers and departments.
• Simulation of the entire system reduces the number of prototypes needed, thereby cutting down on time and costs.
• Automatic generation of software and hardware code from the model reduces the potential for human error.

* The 'model' created with tools like MATLAB/Simulink is often called a 'Moving Specifications' because it represents complex calculations and processes in a Block Diagram, making it highly readable.

As shown in the diagram below, with traditional methods, overall evaluation is not possible until the integration testing stage. If problems are found during the evaluation, the results are sent back to the design phase for further prototyping. Even if the integration testing is successfully passed, any issues discovered during system testing require returning to the design phase with larger scale improvements. The larger the system, the greater the impact on time and cost.

With 'modeling,' as shown in the diagram below, evaluation through simulation becomes possible from the early stages of design. As the design of each part progresses, the accuracy of the evaluation improves, allowing the system to reach completion with fewer prototypes.