Running complex full vehicle integrated co-simulation often requires utilizing virtual ECUs (vECUs) from various sources. Compiling these vECUs as functional mock-up units (FMUs) provides the flexibility to run in different environments. However, ECU software typically contains thousands of calibrations that must be updated as the vehicle is developed. Exposing each calibration as an FMU parameter can make updates tedious in certain toolchains. In this presentation, see how MathWorks and GM collaborated to develop a toolbox that allows an FMU to be exported from Simulink®, enabling it to read a single calibration file directly. This functionality allows users to easily point the FMU to a calibration file during simulation, automatically calibrating it before the simulation runs. This eliminates the need to manually or script updates for thousands of parameters in each environment where the vECU is used.

Effective thermal management in electric vehicles (EVs) is critical for ensuring component safety and passenger comfort. As the demand for operational efficiency in EVs increases, there is a pressing need for advanced thermal management systems that integrate air conditioning, heat pumps, and intelligent control mechanisms. Comprehensive testing and validation of these control algorithms require models that accurately capture the interactions among thermal-fluid components while meeting real-time simulation requirements for hardware-in-the-loop (HIL) testing. The complex dynamics within the refrigeration subsystem, particularly involving AC and heat pumps, often present significant challenges to achieving real-time performance.

In this talk, explore a workflow utilizing reduced-order modeling (ROM) techniques to address these challenges. Through a case study featuring an EV thermal management system with a heat pump, modeled in Simscape Fluids™, see the creation of a data-driven, reduced-order model for the refrigeration subsystem. This model is then integrated with coolant and cabin models, successfully capturing the effects of refrigerants on other working fluids and providing essential feedback signals to the controllers. The resulting model is both real-time capable and sufficiently accurate for real-time HIL testing.

This session aims to equip engineers with the knowledge and tools to efficiently navigate the complexities of real-time simulation in EV thermal management systems. Learn how to apply this workflow and methodology to accelerate simulations of complex thermofluids models within Simscape™ or to incorporate models developed in third-party tools into Simulink® for real-time simulations.

Discover an effective and iterative approach for converting Eaton’s legacy handwritten motor controller software to Model-Based Design using Simulink®. Originally developed a decade ago, the software faced significant challenges, including poor readability, maintainability issues, and extensive use of global variables, making direct reuse difficult. The conversion strategy focuses on systematically transforming software architecture, refining components, and adapting workflows to align with Model-Based Design principles.

A key feature of this process is its flexibility, allowing for gradual conversion and the option to retain portions of handwritten code where necessary. Currently in its fourth iteration, the process benefits from human insights to identify areas for improvement, supported by a robust workflow that facilitates continuous enhancement of both conversion efficiency and code quality. Key elements include modularizing code structures, improving documentation, and establishing a clear workflow for integrating Model-Based Design practices.

In this presentation, get insights into the methodologies and strategies employed. See how this adaptable process can be applied to similar transitions in other organizations, and learn a structured approach to overcoming common challenges in legacy code modernization.

The automotive industry’s increasing reliance on simulations for efficient vehicle development necessitates a robust infrastructure to manage and scale these simulations effectively. This presentation expands upon a key module within Rivian’s Vehicle Simulation Interface tool, previously presented at MAC 2023: a virtual vehicle configurator designed to promote data integrity and consistency. Recognizing that reliable simulations depend on both accurate models and trustworthy input data with full traceability, this configurator focuses on the latter, ensuring confidence in simulation results.

This solution based on MATLAB® and Git™ empowers engineers to construct virtual vehicle data structures by combining modular components such as the aero, springs, motor, battery, tires, and other constituent “parts.” By enforcing the use of gold-source approved data, the configurator guarantees consistency and traceability of vehicle parameters within simulations. Leveraging MATLAB App Designer and object-oriented programming, it provides an intuitive interface for building and managing these configurations. Git integration provides a robust foundation for version control, facilitating collaboration and streamlined data management. The system employs a structured JSON file format for storing component data, promoting interoperability with tools beyond MATLAB. This approach enhances ownership and traceability, linking components to specific engineers and tracking their development lifecycle. Crucially, the configurator accelerates product development by enabling rapid iteration on vehicle components within a simulation environment—a process that would be impractical with physical prototypes in a short timeframe. By automating data management and ensuring data integrity, the system minimizes manual effort, allowing engineers to focus on design and analysis.

In this presentation, explore the configurator’s architecture and see how the synergistic use of MATLAB, Git, and JSON files create a robust and scalable solution for managing vehicle simulation data, ultimately improving efficiency and collaboration in virtual vehicle development.

Electrochemical modeling of batteries, which involves ion and electron transport and kinetic reactions, has traditionally been used to understand how batteries work and optimize cell designs. Recently, these models have gained traction in optimizing fast charging currents, enhancing battery life modeling accuracy and improving battery state estimation. This adoption is accelerated by increased computing power from on-board controllers and the cloud. In this presentation, see the latest advancements in electrochemical modeling capabilities in MATLAB and Simulink, and explore how these products lay the groundwork for engineers to develop high-fidelity electrochemical models—a crucial first step towards better battery control in automotive applications. Please note that the focus will be on electrochemical model development rather than fast charging implementations.

This presentation introduces an architecture-based validation technique aimed at enhancing the model-based development of embedded systems. It emphasizes the importance of systematically measuring the “goodness” of individual components and overall system architecture to achieve targeted outcomes using metrics as a feedback mechanism. This presentation covers various metrics for evaluating software architecture, comparing design choices, and highlighting trade-offs with practical examples. With a focus on reducing cruft and technical debt and ensuring high-quality, cost-effective real-time systems, this technique uses System Composer™ as a tool for architecture analysis, enabling temporal checks, system simulation, and the generation of key artifacts like UML diagrams. This ensures that local design choices align with global system behavior, fulfilling features, requirements, and architectural constraints and offering insights into model-driven approaches for developing robust embedded software architectures.

Explore the integration of MATLAB^® and Simulink^® for AI with Boston Dynamics’ agile robotic platform, SPOT. You will see both a presentation and a demonstration showcasing the capabilities of performing predictable autonomous inspections. Learn the initial steps for conducting industrial inspections using SPOT, a robust robot equipped with professional-grade sensors. These sensors feed data into MATLAB, which serves as a platform for the development and deployment of complex AI algorithms.

Nikola is a class-8 truck OEM with battery and hydrogen fuel cell electric trucks. These products have multiple ECUs that communicate with components, supplier ECUs, and amongst themselves. One such ECU, the vehicle control module (VCM), is responsible for varied vehicle functions and is tested extensively with hardware-in-the-loop (HIL) via regression testing. Log analysis of these tests to provide a pass/fail is the most critical task of regression testing. A setup was created using the MATLAB® unit test framework, Simulink Real-Time™, and Simulink Test™ to run hundreds of tests with a single command. The automation setup also includes the use of Vehicle Network Toolbox™ to assess results from log files of each run. Lastly, results are exported to Excel files with test details and results using MATLAB tables. All of this is accomplished with one test harness build, i.e., a universal test harness. The automation workflow decouples an engineer’s time from regression testing, and they can only analyze failures after the entire suite is tested. It also allows scheduling test execution and analysis at night, so the engineer can focus on the development of new tests using the same hardware during the day. Therefore, the setup has had a positive side-effect where the HIL apparatus is used at capacity without much idle time. In this session, you will get multiple code examples to inspire you to use such an approach in your own workflows, and learn:

• The traditional way of HIL testing (multiple harnesses and builds) and its disadvantages
• The use of Simulink Real-Time API to create test scripts
• The use of Vehicle Network Toolbox to assess log files
• How to create test classes with unit test framework in MATLAB
• How to integrate multiple test classes to be run with a single command
• The advantages and potential pitfalls of the current process
• The potential solutions to avoid these pitfalls

In this presentation, explore the development of automotive applications using cloud-to-cloud simulation and reprocessing techniques featuring RoadRunner and Simulink®. Learn how these advanced methods can enhance the efficiency and accuracy of automotive software development toolchains, leading to more reliable and innovative automotive solutions. Key topics include the integration of cloud-based platforms, the benefits of cloud-to-cloud communication, and case studies demonstrating successful implementations in the industry.

Automated driving testing requires accurate and precise road representation to ensure reliable simulations. In this presentation, get an overview of workflows that integrate GeoMate HD maps into RoadRunner, facilitating the creation of high-definition scenes for automated driving (AD) testing. GeoMate’s RealSimE (real-world simulated environment) maps, generated using high-resolution satellite and aerial imagery powered by AI-driven mapping technology, are seamlessly integrated with RoadRunner to enable the design of realistic simulation environments. These environments include detailed road networks, intersections, and traffic features, supporting AD and ADAS testing for tasks like path planning, localization, and edge case analysis. By providing scalable, cost-effective, and high-accuracy mapping solutions, GeoMate enhances the realism and efficiency of simulations, advancing the development of safe autonomous vehicles.

Vehicle simulations are used in varying capacity during the design and development of automobiles. Besides capturing the vehicle physics, these models often require the ability to include control algorithms, driver behavior, and environmental factors. With this complexity, models need to be parameterized and validated against real vehicle data to be able to use them with accuracy and confidence.

To streamline this process, FEV and MathWorks collaborated on developing a vehicle model using Powertrain Blockset™, parameterizing it using FEV’s extensive benchmarking data and then comparing the results to the vehicle data collected from a chassis dynamometer. Learn how to build vehicle models using the Virtual Vehicle Composer app, and the steps required to parameterize the model and get the results within acceptable accuracy to enable you to use the vehicle models in your workflows.

In-cabin safety has been under investigation in recent years. Relevant predevelopment research could be complicated due to unclear requirements, and therefore become data dependent. The abundant toolbox resources in MATLAB® are used by AANA engineering teams to conduct research on in-cabin sensing projects, e.g., occupant classification and in-cabin localization for removable equipment. In this session, AANA shares how using MATLAB could help accelerate the in-cabin engineering research, specifically when neural network techniques are involved. The Deep Network Designer app (from Deep Learning Toolbox™) makes the research process painless when dealing with large amounts of data.

To conduct chassis controls testing through simulation, several components are essential. Engineers need to select the type of vehicle for testing their chassis control algorithms. They may have data to parameterize the vehicle model according to specific requirements, or they might only know the general vehicle class. Additionally, they must choose the types of tests to perform—whether simulating standardized scenarios or creating custom ones. Defining the scenario’s environment, including terrain and friction properties, is also crucial. These elements must integrate into a simulation platform to execute the tests and generate reports.

In this presentation, see MathWorks engineers demonstrate how scene and scenario information defined in RoadRunner can integrate with a physics model defined in Simulink®. RoadRunner will be used to create a 3D scene and scenario based on the AVL ZalaZONE Automotive Proving Grounds in Hungary. The Vehicle Dynamics Blockset™ will define a closed-loop vehicle model. This workflow will be showcased using a traction control system to test vehicle performance on a split-mu surface while climbing a hill.

The rapid adoption of electric vehicles and the need to adhere to warranty statements require robust diagnostics and prognostics of battery state of health. In this talk, learnan approach to battery degradation diagnosis, focusing on capacity fade estimation. Explore the integration of Least Squares (LS) and Kalman Filter algorithms, deployed on a virtual electronic control unit (ECU) as a service, to provide real-time insights into battery health.

The SOH algorithm is deployed on a virtual ECU. The vECU is deployed on the cloud while still allowing closed-loop simulation with a plant based  on Simulink®. This virtual deployment allows for extensive testing and validation without the need for hardware prototypes, significantly reducing development time and cost. The virtual ECU works using Adaptive AUTOSAR® middleware and can be tested in desktop simulation or hardware-in-the-loop.

A new trend with automotive OEM’s adopting software-defined vehicle processes has been utilizing cloud-based virtual desktops and infrastructure to develop and test models and software. Join AWS to learn more about how you can utilize virtual engineering workbenches to accelerate your team’s Model-Based Design with MATLAB®, Simulink®, and other partner offerings to increase collaboration, reduce startup time, and collaborate more effectively across your organization.

In this presentation, discover how to achieve automated driving through urban intersections by developing key components such as lane change management, traffic signal compliance, and collision avoidance. Learn about vehicle localization using RoadRunner HD maps, and discover a traffic signal follower and collision avoidance system through V2X communication. Additionally, see the design of a motion planner to generate optimal trajectories that avoid dynamic bottlenecks caused by parked vehicles.

This session is tailored for engineers involved in designing onboard battery chargers and their integration into the overall vehicle system. You will explore system-level design approaches for thorough validation of system operations, fault analysis, and safety assessments. Additionally, hear about the validation of and interaction with external systems, equipping you with the tools to optimize your design. This includes architecture decisions, such as determining the optimal placement of software—whether on the actuator or within a central compute unit.

Highlights

• Behavioral charger models for fast battery charging simulation         
• System-level design and analysis of battery chargers
• Debugging system interactions
• Fault analysis

As software-defined vehicles continue to evolve, Vector and MathWorks are collaborating to enhance AUTOSAR® Adaptive and Classic workflows, enabling more efficient development, validation, and deployment of automotive applications. This session will showcase the vehicle control system (ABS for AUTOSAR Classic Platform), fully modeled in Simulink® and integrated with MICROSAR Classic, performing virtual ECU simulation with AUTOSAR Blockset and Vector software tools. Additionally, we will explore how the integration of MICROSAR Adaptive and Simulink can streamline the development of AUTOSAR Adaptive applications and enhance cross-tool compatibility.

Key topics covered include:

• How Vector tools and MATLAB and Simulink are utilized at different virtual ECU testing levels
• Future opportunities for enhancing compatibility and automation in SDV software workflows