Software-Defined Vehicles: Developing Service-Oriented Applications with Simulink

1:00–1:20 p.m.

Our driving experience will soon be defined by the software running in the car and in the cloud. A new holistic approach for software architectures based on services and service-oriented communication is emerging. This approach enables continuous development and deployment of innovative software features and makes new types of collaboration possible between OEMs and software platform providers.

Join this session to learn how you can use Simulink® to design, simulate, and deploy applications into these new software architectures according to industry standards such as AUTOSAR Adaptive, DDS, and ROS.

Luigi Milia

Luigi Milia, MathWorks

Shwetha Bhadravathi Patil

Automotive DevOps for Model-Based Design with AWS, NXP, and MathWorks

1:20–1:40 p.m.

Vehicle features and capabilities are transitioning from being mostly mechanically defined to being software defined. OEMs are adopting agile software development methods to update and maintain this software, driving the need for DevOps. NXP Semiconductors, MathWorks, and AWS have collaborated on a DevOps solution for Model-Based Design utilizing advanced vehicle control algorithms. In this talk, NXP will present the full cloud-to-vehicle solution built with AMS CodeSuite services by using MathWorks design tools targeting NXP automotive processors. This presentation also covers NXP® Model-Based Design Toolbox supporting code execution and profiling on the NXP S32S GreenBox II and NXP S32G GoldBox.

Curt Hillier

Curt Hillier,
NXP Semiconductor


Model-Based Design Meets CI: I Connected Simulink with My CI System—What’s Next?

1:40–2:00 p.m.

Many development teams have started integrating Model-Based Design into their continuous integration (CI) pipeline. What’s next? Learn how to optimize the interactive development practices to increase CI flow rate, maximize automation in the pipeline, and plan the extension of CI to DevOps.

Tjorben Gross, MathWorks


Challenges and Successes in Migrating Legacy Software Modeling to a Simulation-Based Product Development System

2:00–2:20 p.m.

To meet the demands of the modern on-highway market, engineering teams are moving towards agile development with a continuous integration and delivery development methodology. To adopt this new methodology, sometimes evolutionary changes in product and process aren’t enough and a revolutionary approach must be embraced, including knowing which parts of 100 years of development are integral to the company and which parts are holding it back from growth. Cummins has been reshaping its product and process to meet these challenges, targeting more effort on simulation and architecture with an initiative called simulation-based product development. MathWorks products and services are a strategic part of that vision. We will discuss how key enablers like executive sponsorship and vision, AUTOSAR, CICD, and “simulate everything” are helping to cement change and surmount challenges in culture, legacy technical debt, and skill sets.

Jason Stallard, Cummins

Brandon Trombley

Brandon Trombley, MathWorks


Turning the Tables on Validation with Agile Model-Based Design

2:20–2:40 p.m.

The growing complexity of today’s products—fueled by ever-increasing software content and coupled with shrinking timelines for delivering new products—requires engineers to “turn the tables” on existing workflows. The core principle of agile Model-Based Design centers on the idea of reducing rework. Typical workflows result in on-product validation at the end of the development cycle. Requirement issues found at that point drive rework—often taking the design through the entire development cycle again. Early integration and validation through agile Model-Based Design provides an opportunity to minimize rework through up-front, rapid iteration on architecture and requirements.

Jim Ross

Jim Ross, MathWorks


Building the Digital Thread from MBD to MBSE to Meet ISO 26262 for Embedded Software

2:40–3:00 p.m.

Adoption of ISO 26262 has led to the expansion of development methodology from hand code to Model-Based Design. A number of process improvements were identified to achieve traceability and thread pulling across the system engineering process, including connecting architecture models and implementation models, implementing traceability between requirements and models, and understanding model architecture’s implication on impact analysis. This presentation will provide an early overview of the solution that involves an integrated Model-Based Design-Model-Based Systems Engineering workflow, limited duplication of sources of truth, establishment of traceability and coverage with a requirement management tool, and a componentized modeling style.

Joshua McCready,
Ford Motor Company

Hans Gangwar,
Ford Motor Company


Addressing Common Inefficiencies When Targeting AUTOSAR and ISO 26262 with Simulink

3:00–3:20 p.m.

Although the AUTOSAR standard has included constructs to support functional safety concepts since its inception, the automotive functional safety standard ISO 26262 came later from a different community. As a result, an engineering team needs to understand the gaps between these two standards when applying both at the same time. If done appropriately, the team can reduce the effort required for achieving compliance. However, finding the right solutions could be time consuming and done at the expense of engineering time on the product. In this presentation, explore a set of best practices in AUTOSAR for adhering to ISO 26262, distilled from working with automotive engineers in the last 5 years, covering architecture, data management and transfer, and workflow.

Michael Boyle

Michael Boyle, MathWorks


Effective Model-Based Development Strategies for ASPICE and Safety Compliance

3:20–3:40 p.m.

Overcoming the challenges of effective software architecting and detailed designing in hybrid model-based and code-based environments, and doing so in a compliant manner, requires strategic decisions among the project team. These strategies, as well as compliance, are often not well formed or understood by project teams, and the answers are complicated in hybrid environments. This session will demonstrate how model-based development can be compliant in similar fashion to hand code but also have distinct advantages, especially when utilizing key features of Simulink®.

Peter Abowd,
Kugler Maag Cie

Steve Tengler,
Kugler Maag Cie


Faster Software Delivery with Polyspace in Your Software Factory

3:40–4:00 p.m.

Increasing C/C++ software content and faster delivery requirements are common trends across vehicle software development organizations. Static analysis is a critical technique for verifying that the code meets quality, safety, and security goals. Learn about the benefits of using static code analysis at each step of the software development workflow, beginning with coding and during code integration. See how Polyspace® code verification tools can be integrated into developers’ IDE and software factories to enable faster software delivery and higher quality code.

Patrick Munier

Patrick Munier, MathWorks


Building a Cloud-Based Digital Twin for an EV Battery Pack

4:00–4:20 p.m.

Creating, validating, and correlating the model of a physical asset is important to building a digital twin, but modeling is only one aspect of the overall process of developing and deploying digital twins. In this presentation, you’ll explore a project which spans from developing the model of an EV battery, deploying it to the cloud and connecting it to the data infrastructure, and predicting battery state of health based on data from a real-world electric vehicle fleet. You’ll also learn about key considerations when planning your digital twin project.

Will Wilson

Will Wilson, MathWorks

Battery SOH and SOC Estimation Using a Hybrid Machine Learning Approach

12:00–12:15 p.m.

KPIT developed a hybrid approach to overcome the shortcomings of existing individual methods for SOC and SOH estimation. It combines a battery model and a neural network to predict SOC and then uses the obtained SOC to derive SOH parameters. Deep Learning Toolbox™ and MATLAB® were used to train a feedforward neural network, which was then extensively validated for robustness. The neural network was then incorporated into Simulink® and deployed to a PowerPC-based embedded platform using Embedded Coder® and AUTOSAR Blockset. This workflow has been validated on multiple datasets for LFP and LCO chemistry. It has provided SOC and SOH estimation with improved accuracy well within +- 5% consistently over a different driving cycle range.

Mahesh Ghivari

Mahesh Ghivari,
KPIT Technologies Limited

Debango Chakraborty

Debango Chakraborty,
KPIT Technologies Limited


Developing Onboard SOH Estimation Using DVA and ICA for LFP Batteries

12:15–12:30 p.m.

In this session, see how we developed a high accuracy onboard battery state of health estimation method based on the differential voltage (DVA) and incremental capacity analysis (ICA) for electric vehicles. Using cycling data from lithium-ion battery cells at various temperatures, we extracted the charging cycles and calculated the DVA and ICA curves, which are then filtered with an IIR-filter to reduce noise. Multiple features (i.e., peaks or valleys of the curves) are extracted and analyzed, and the most promising features are selected for further steps. The selected features are brought into correlation with the capacity fade, and a linear regression model is calculated between the selected features at various temperatures. With these linear models, a 2D Look-Up Table (LUT) is created by interpolating the values between the linear models. For the onboard implementation, we developed a Simulink® model which realized the calculation and filtering of the ICA- and DVA-curve. Also, we implemented a feature detection algorithm that detects and verifies the selected features, which are forwarded to the 2D LUT to calculate the current SOH. We tested and converted this model to AUTOSAR-compliant code and will validate it on Gotion’s in-house developed BMS.

David Jaurnig

David Jauernig, Gotion, Inc.


Onboard Battery Pack State of Charge Estimation Using a Trained Neural Network

12:30–12:45 p.m.

Using battery cell charging data stored in Gotion’s cloud data platform, we train and validate a neural network to estimate pack state of charge (SOC) during vehicle charging with the Deep Learning Toolbox™ and in-house data query API. We create an onboard SOC estimation strategy in Simulink® using the trained neural network. Afterward, we verify the algorithm’s ability to meet functional requirements using Simulink Test™ and deploy it as a “shadow” strategy within an existing AUTOSAR SOC software component. The impact on CPU and memory resources of the microcontroller is evaluated first. Then, we evaluate the neural network-based SOC estimator on test vehicles and find that the results (< 3%) are promising.

Trevor Jones

Trevor Jones, Gotion, Inc.


AI Workflows for Battery State Estimation

12:45–1:00 p.m.

State of charge (SOC) estimation is among the most important tasks of a battery management system (BMS). SOC estimation is typically performed by current integration or using a Kalman filter. In this session, we will describe an alternative method based on AI. A deep neural network is trained to predict SOC based on voltage, current, and temperature measurements. The resulting network is then implemented in Simulink® and incorporated into a closed-loop BMS model. Finally, C code is automatically generated from the net for hardware implementation on an NXP S32K3 board used in PIL mode.

Javier Gazzarri

Javier Gazzarri, MathWorks


Building a Virtual Vehicle for Large-Scale Simulation Studies

1:30–1:50 p.m.

The importance of virtual vehicle simulation has only grown as the automotive industry shifts to a more electrified, autonomous, and connected world. Their impact across organizations, however, is limited by the challenge of building models that can answer engineering questions with acceptable accuracy and speed. Building the vehicle model is only the start, and the platform must support the need to scale up the simulation studies needed for engineering decision making. Discover recent developments at MathWorks that make the process of building a flexible, customizable simulation platform easier and more automated. See an example showing how the virtual vehicle models can be deployed to the cloud for large-scale simulation studies.

Mike Sasena

Mike Sasena, MathWorks

Brad Hieb

Brad Hieb, MathWorks

Scott Furry

Scott Furry, MathWorks


Applying AI Technologies to Vehicle Sensor Modeling

1:50–2:10 p.m.

This presentation introduces a method of applying AI technology to vehicle sensor modeling. In this method, machine learning with non-parametric regression is applied. A detailed GT-SUITE vehicle model is coupled with Simulink® to generate design data needed for sensor modeling. Relevant predictors are selected. The AI training algorithm is applied. The NOx sensor model is designed based on the data generated from the detailed GT-SUITE vehicle model. The FTP-75, US06, and HWFET are used for the vehicle running setup. The developed NOx sensor model is embedded into Simulink and coupled with the GT vehicle model to verify the prediction capacity. The LA92 driving cycle and Brazilian road test are used on the sensor model validation. The machine learning technology on NOx sensor modeling proves to be successful and will have a wide range of applications in the automotive industry.

Rafael Átila Silva, Stellantis


Multi-Stack Fuel Cell Electric Vehicle Modeling and Applications

2:10–2:30 p.m.

Multi-stack fuel cells offer various performance improvements over traditional fuel cell systems. In this talk, see how MATLAB® and Simulink®  can be used to simulate these systems on a component and system level. As an application of this type of model, investigate a control approach to improve the overall efficiency of this multi-stack electric vehicle model for a given drive cycle.

Jason Rodgers

Jason Rodgers, MathWorks


What’s New in MATLAB, Simulink, and RoadRunner for Automated Driving Development

2:30–2:50 p.m.

MATLAB®, Simulink®, and RoadRunner help engineers to build automated driving systems with increasing levels of automation. In this session, you will discover new features and examples in R2021b and R2022a that will allow you to:  

  • Author scenes and scenarios for driving simulation 
  • Simulate sensors and vehicle dynamics 
  • Design detection, localization, sensor fusion, planning, and controls algorithms 
  • Deploy to C, C++, GPU, and ROS
  • Test functionality and code
Pitambar Dayal

Pitambar Dayal, MathWorks


Converting Spreadsheet-Based Scenario Definitions to OpenSCENARIO Files

2:50–3:10 p.m.

The goal of this project was to create a tool to convert spreadsheet-based scenario definitions into the OpenSCENARIO standard format for easy exchange, usability, and storage for Ford engineers. Through standardization, the tool allowed for scenario reuse regardless of the simulation tool and test case format. The tool was first used to support automated conversion of scenarios defined in complicated spreadsheets to the beta version of OpenSCENARIO. This tool is capable of showing the resulting scenario in the Automated Driving Toolbox™ drivingScenario tool and exporting OpenSCENARIO 0.9.1 files representing the corresponding CarSim scenarios. Lastly, the app was modified so that users could enter information required by OpenSCENARIO to build up scenarios from other documentation—enabling reusability and easy shareability across Ford.

Emily Foster

Emily Foster,
Ford Motor Company


Design and Simulate Scenarios for Automated Driving Applications

3:10–3:30 p.m.

The development of advanced driver-assistance systems (ADAS) and autonomous driving applications often depends on simulation to reduce in-vehicle testing. The automotive industry is investing in standards like OpenSCENARIO to describe dynamic content in these driving simulation environments. In this session, you will learn how to author scenarios on realistic road networks designed in RoadRunner. You can use this workflow to simulate scenarios with built-in agents as well as author and integrate custom agents designed in MATLAB®, Simulink®, or CARLA. The scenarios can be exported to OpenSCENARIO for simulation and analysis in external tools if desired. Using this workflow, you can quickly author and simulate scenarios to gain system insight and test your designs.

Shusen Zhang

Shusen Zhang, MathWorks


From Motorcycle to Chevy Bolt: A Journey with MATLAB in Autonomous Vehicles and Robots Research

3:30–3:50 p.m.

Hear about our two decades of lessons learned and experience with autonomous vehicle and robot navigation through four case studies in chronical order. We started our journey by developing the world’s first autonomous motorcycle to participate in the DARPA Grand Challenges (DGCs) 2004 and 2005. We developed perception, vehicle navigation, and control algorithms for this non-minimum phase system to navigate in a desert terrain. Following DGC, we developed intelligent navigation algorithms for skid-steered vehicles. Collaborating with Johnson Space Center, we have also developed localization and mapping algorithms for NASA Robonaut, a humanoid robot serving on the International Space Station. Most recently, our students have successfully competed in the first GM/SAE Autodrive Challenge. MATLAB® has been an irreplaceable toolset and an integral part of our developing experience.

Dezhen Song,
Texas A&M University


Design of a Vehicle Platooning Controller with V2V Communication

3:50–4:10 p.m.

Learn how to design a controller for vehicle platooning applications with vehicle-to-vehicle (V2V) communication. Every following vehicle in a platoon maintains a constant spacing from its preceding vehicle. Vehicles traveling in tightly spaced platoons can improve traffic flow, safety, and fuel economy. Each vehicle obtains the position and movement information of the other vehicles in the platoon wirelessly via the V2V communication. A given acceleration profile drives the lead vehicle, and every trailing vehicle follows the lead vehicle while maintaining a predefined space by a platooning controller.

Seo Wook Park

Seo-Wook Park, MathWorks