Model-Based Design and Automation Streamline Hybrid Vehicle Software Development
Standardized Development and Verification Environment Accelerates Vehicle Control Software Development
Partner with Us for Success
Get expert guidance from MathWorks Consulting on tailored solutions to help you achieve your goals.
As one of the world’s leading automakers, Hyundai Motor Group develops sophisticated drive control software that is central to vehicle performance and functionality. Historically, software development and verification have relied on a sequential, hardware-centric process that limits flexibility and slows iteration. By adopting a more integrated, model-based workflow with MATLAB® and Simulink®, Hyundai standardized its development and verification process, enabling engineers to simulate, validate, and test software earlier in the development cycle while establishing a foundation for future software-defined vehicle (SDV) initiatives.
Drive control software modulates and coordinates a variety of vehicular functions, including fuel economy, safety features, and motion dynamics. Verifying the software’s functionality and safety is critical, and each electronic system in a road vehicle must comply with ISO® 26262 for functional safety. Working with MathWorks Consulting Services, the Hyundai team created a unified development and verification environment, giving engineers access to modeling, simulation, code generation, and verification capabilities within a common workflow.
Hardware-Dependent Verification and Growing Pains
Developing drive control software for road vehicles that adheres to international standards is no small task. After generating code, Hyundai engineers perform static verification against applicable coding rules. These checks are one part of the broader development and verification activities required to support ISO 26262 functional safety objectives.
“In other words, it is to make sure that the software that you deploy and produce is safe and that it does exactly what it says it should do,” says Ren Sang Nah, a senior principal technical consultant at MathWorks.
For Hyundai, developing the appropriate software and meeting these requirements was a cumbersome process. “Previously, developing and verifying software required actual automotive hardware,” says Jin Ae, a researcher on the vehicle drive control development team with Hyundai Motor Group. Much of the verification and validation (V&V) work was oriented around physical hardware, and many software tests were performed directly in the vehicle. Because developers also executed much of the process manually using a variety of tools, verification could take longer and was difficult to begin before the hardware was available.
Further, not all developers had expertise in every tool required for the workflow. The variety of development and verification tools created a learning curve and made it more difficult for every developer to execute the workflow consistently.
As a result, verification occurred late in the process and depended on hardware testing; engineers could discover issues only after testing the software in a vehicle. They would then need to revise the model and return to vehicle testing, extending the development and verification cycle.
Ae’s team used a V-model workflow for software development, in which each phase of verifying and validating software follows a linear, sequential path. To keep up with the pace of SDV development, Hyundai needed to shorten the verification cycle by establishing an environment that was less dependent on physical hardware.
Model-Based Design for Efficient Verification Workflow
Developers and researchers at Hyundai Motor Group began looking for a way to simplify the workflow and make development and verification tools easier for engineers to use. After exploring their options, Model-Based Design with MATLAB and Simulink emerged as an answer.
“With this project, developers can now simulate and execute validation easier than before. We can reduce the time to verification and reduce the time spent in the development stage. We also reduce some of the costs.”
Hyundai began introducing the new development and verification process within its vehicle control software programs. Through a multiyear engagement with MathWorks Consulting Services, Hyundai deployed and customized MBD Workbench, a Model-Based Design workflow framework, to complement the MATLAB and Simulink toolchain and support Hyundai’s process. Ae’s hybrid vehicle group subsequently integrated the environment into its development process and adapted it to the group’s workflow.
Ae’s team begins by analyzing requirements for new or enhanced drive control functions. Engineers extract and review the software requirements and then use Simulink to develop the corresponding models. Desktop simulation enables them to evaluate software behavior earlier in the development process.
Once the design is validated, the team uses Simulink and Embedded Coder® to generate production code for vehicle controllers. Throughout development, engineers use MATLAB and Simulink to verify models and code and to generate the documentation needed for ISO 26262 certification activities.
The workflow extends beyond the development team. Vehicle-level testing uses Simulink for software-in-the-loop and hardware-in-the-loop testing, while a plant model developed with Simscape™ supports validation simulation before testing in the actual vehicle.
Beyond automation, one of the key advantages of using MATLAB and Simulink is that it improves the whole development and verification workflow. “Another reason that we adopted Model-Based Design in MATLAB was to improve the usability of software,” said Ae. Previously, the team required a wide variety of tools. MBD Workbench brings development and verification capabilities together in an integrated workflow framework, making them easier to access and use.
“With this project, developers can now simulate and execute validation easier than before,” he says. “We can reduce the time to verification and reduce the time spent in the development stage. We also reduce some of the costs.”
Standardizing and Automating Verification
With the new workflow, Hyundai can perform more development and verification activities through simulation before testing the software in a vehicle. This shift-left approach helps engineers identify issues earlier in the process and increases the team’s internal verification capability. For example, improving fuel economy requires calibrating numerous parameters in the drive control software. Previously, engineers often cycled between software development and vehicle testing, with defects discovered during hardware testing triggering costly rework. By validating software earlier with MATLAB and Simulink, the team can reduce the number of issues that reach vehicle testing and accelerate development.
“We were impressed by the way the MathWorks consultants combined technical expertise with an understanding of our workflow. Their support was not limited to implementing the tools—they also helped us structure the process and build internal capability so that our team could continue improving it independently.”
This paid off during a recent project to improve fuel economy software. Calibrating these functions previously required extensive work with a real vehicle and hardware because of the number of calibration parameters involved. “Now, because we are using simulations, this calibration has become much more efficient,” says Ae. “We were able to sign off on the prototype model relatively quickly and without issues.”
The new process also facilitates continuous integration, and the team’s next step is to automate model-level testing as well, including generating test cases. This way, they can find and fix defects faster. “We can execute this manually, but it could cost many human errors,” says Ae.
MathWorks Consulting Services supported Hyundai in establishing the new workflow. Hyundai brought its development and verification requirements to the consultants with the goal of making the tools easier for developers to access and use. Together, they developed an environment customized to Hyundai’s process.
“We were impressed by the way the MathWorks consultants combined technical expertise with an understanding of our workflow,” says Ae. “Their support was not limited to implementing the tools—they also helped us structure the process and build internal capability so that our team could continue improving it independently.”
This support was instrumental in smoothly incorporating the new tools and workflow. “In our organization, we have found that MathWorks tools and consulting services have been helpful in improving efficiency in areas such as development and verification,” he says.
Supporting Faster Software Release Cycles
Software-defined vehicle initiatives are increasing the need for shorter software release cycles and rapid verification. For Hyundai, that need has influenced the development of a verification platform that can support both current controller development and future vehicle architectures.
“Changing to software-defined vehicles is the first main goal of our group,” says Ae. “And the second main goal is to reduce the cost and time on verification.” Improving the verification process helps the group respond to the shorter release cycles associated with that transition.
Rather than waiting until hardware testing to perform much of the verification, Hyundai can now move more simulation and validation activity earlier. Automation connects code generation with static verification and software builds, while the planned expansion into model-level test automation will further reduce manual work.
By integrating more verification into the development process, Hyundai is establishing a more consistent workflow for functions related to fuel economy, actuator control, and vehicle safety. Developers can create simulations for their own functions, use measured vehicle data to develop test cases, and evaluate changes before testing in the vehicle. This standardized approach improves the efficiency of model V&V today while supporting the faster development cycles required for future software platforms. “We are proceeding with development in a way that allows us to immediately leverage the systems and processes our organization has already adopted and utilized,” says Ae. “We feel more confident in our processes and the quality of the software.”
Partner with Us for Success
Get expert guidance from MathWorks Consulting on tailored solutions to help you achieve your goals.
Read Other Stories
Johnson Electric Uses Digital Twins to Validate Novel Actuator Design and Control Strategies
Reducing Validation Time for Sensorless Motor Control Before Physical Prototypes
Iveco Group Accelerates Electric Vehicle Software Development with Virtual Vehicle Simulation
Fully Virtual Vehicles Enable Faster, More Scalable EV Software Testing
Transitioning Legacy Motor Controller Software Development to Model-Based Design
Eaton Adopts a Center-Out Strategy to Convert Legacy Motor Control Software