Model Dr. MDiA
SUBARU CORPORATION
We present a case study on how adopting and leveraging Model Dr. MDiA reduced model operation and management man-hours, thereby improving development efficiency.
Significant Reduction in Model Review Man-Hours
In recent years, automotive systems have become increasingly large-scale and complex. In Model-Based Development as well, to ensure the quality of these complex models and effectively operate and manage them, companies need to build model-centric toolchains tailored to their actual operations.
To address this challenge, Ryomo Systems developed "ModelDr.MDiA," a JMAAB control modeling guideline checking tool. The goal is to support efficient product development (e.g., reducing model operation and management man-hours) by providing a mechanism to check modeling rules tailored to each customer.
This case study demonstrates how SUBARU CORPORATION, which drives the development of EyeSight control software, reduced model operation and management man-hours and improved development efficiency by adopting and utilizing ModelDr.MDiA.
Department Overview
SUBARU CORPORATION develops EyeSight, an advanced driver assistance system centered around stereo cameras. EyeSight is equipped with a wide variety of features, including Adaptive Cruise Control with Full-Speed Range, Pre-Collision Braking, Lane Departure Prevention, and Lane Centering Control. The basic architecture of EyeSight consists of two main components: the "Image Processing Unit," which processes images captured by two stereo cameras, and the "Control Processing Unit," which executes vehicle control based on vehicle data in addition to that sensor information. In developing the "Control Processing Unit," SUBARU introduced Model-Based Development to rapidly verify specification validity using simulation during the upstream development phase.
Background and Challenges
Prior to the introduction of Model Dr. MDiA, issues existed where model modeling conventions—such as signal line names and subsystem partitioning—were not standardized overall. This resulted in problems such as "increased man-hours for model merging, integration, and review" and "difficulty in making adjustments during personnel shortages due to model reliance on individual engineers."
To resolve these issues, it was necessary to address Challenges 1 and 2 below.
Challenge 1: Standardizing Model Modeling Conventions
Formulate modeling guidelines combining the JMAAB Control Modeling Guidelines, which are essential in the automotive industry, with proprietary rules specific to the adopting department, and apply them to newly created models. By applying these guidelines, SUBARU aimed to resolve the problems of "increased man-hours for model merging, integration, and review" and "difficulty in making adjustments during personnel shortages due to model reliance on individual engineers."
Challenge 2: Standardizing Model Checking and Verification Methods
Standardize both the methods for checking compliance with the established guidelines and the reporting format of check results. Through this, SUBARU worked to mitigate the problem of "increased man-hours for model merging, integration, and review."
Key Adoption Points for Model Dr. MDiA
Point 1: Customization of Diagnostic Items
・ Capability to incorporate proprietary rules into Model Dr. MDiA
・ Capability to group necessary check items
Point 2: Outsourcing Development of Proprietary Rules
・ Ability to outsource the creation of scripts for checking rules
・ Ability to outsource the integration of scripts into Model Dr. MDiA
Point 3: Report Output Functionality
・ Ability to select from multiple formats, such as HTML and CSV, using the standard report output feature to suit operational needs
Benefits of Adoption
Standardized Model Conventions Boost Product Development Efficiency!
Models now comply with guidelines, keeping model quality consistent without depending on individual engineers' skills. This reduced reliance on specific personnel for models, making model merging and integration easier while also facilitating the division of work.
Self-Checks Boost Quality Verification Efficiency!
By having each engineer perform self-checks on models using Model Dr. MDiA from the early stages of development, "model quality was standardized" and "standardized reports were retained." Consequently, model quality verification progressed smoothly, leading to a significant reduction in review man-hours.
Customer Comments
Regarding model quality, there were aspects that previously depended on individual engineers' skills. However, using ModelDr.MDiA allows us to maintain consistent quality for outputs (e.g., models, reports), making it easy for management to review the results. Furthermore, having each engineer identify and correct guideline violations contributed to skill development among our engineers, which has also been a major benefit of this implementation.
Going forward, we aim to further improve efficiency by building a model-centric toolchain and allocate those saved man-hours to product development, thereby achieving both enhanced product competitiveness and quality assurance. (Mr. Kato, Advanced Safety Design Dept., SUBARU CORPORATION)
*Note: Model Dr. MDiA is for use in Japan only (cannot be used overseas).