Virtual Vehicle: Driving Simulator

This demo shows a virtual vehicle model running a driver-in-the-loop test. Join us at the demo booth and see if you can beat the high score.

The highlights include:

• Creating racing-circuit scenes semi-automatically.
• Interactively racing on the generated scene using a hardware steering wheel.
• Generating minimum curvature racing lines.
• Estimating lap time based on vehicle and driver parameters.
• Demonstrating a cyberattack on power-steering (security).

Fast and High-Quality Software Delivery for Software-Defined Vehicles

The development of automotive systems and software is being redefined to deliver perpetually upgradeable software-defined vehicles. Cybersecurity considerations and requirements from ISO/SAE 21434 are a motivating factor to increase the development speed and automate the whole process.

This demo shows how you can leverage cloud resources for scaling software development with virtual analysis and automate testing. This also includes continuous integration for improved collaboration and quality, as well as a virtual processor-in-the-loop to reduce hardware dependency of testing.

The presented workflow is based on a collaboration of Elektrobit, AWS, and MathWorks.

Processor-in-the-Loop Testing of an AI-Based Battery State of Charge Estimation Model

Battery state of charge (SOC) is a critical signal for a battery management system (BMS) but it cannot be directly measured. Virtual sensor modeling can help in situations when the signal of interest cannot be measured, or when a physical sensor adds too much cost and complexity to the design. Deep learning and machine learning techniques can be used as alternatives or supplements to other virtual sensing techniques.

In this demo, learn how to develop AI-based virtual sensor models for BMS SOC estimation using different machine learning techniques, and how to integrate them into Model-Based Design for system-level test and implementation on an Infineon AURIX™ TC4x board via automatic code generation.

Highlights include:

• Designing and training machine learning components using MATLAB^®.
• Importing trained AI models from other frameworks (e.g., TensorFlow) into MATLAB.
• Integrating machine learning models into Simulink^® for closed-loop or end-to-end system-level simulation.
• Enabling easy access to Infineon AURIX™ TC4X through target-optimized automatic code generation.
• Integrated verification via processor-in-the-loop tests.

Autonomous and Automated Driving Development: Perception to Code Generation

Advance the design of automated driving perception, planning, and control systems by enabling engineers to gain insight into real-world behavior, reduce vehicle testing, and verify the functionality of embedded software. Realistic virtual environments are a vital component in the workflow of advanced driver-assistance system and automated driving (ADAS/AD) function development.

Learn and discuss how to create 3D scenes and scenarios with the RoadRunner suite to be used for simulations. Leverage HD map providers via RoadRunner Scene Builder and cosimulate with MATLAB^® and Simulink^® for algorithm development. Learn how to use AI to develop novel perception algorithms or how you can design tracking, fusion, planning, and control algorithms. Finally, generate C/C++ or target specific code and deploy it to your hardware in the vehicle.

System Modeling and Simulation of 4D Imaging Automotive Radar

Imaging radar is considered a key automated driving and safety technology providing robust, accurate, and long-range perception capabilities. This demo shows how to model and simulate a 4D imaging MIMO radar using MATLAB^®. The waveform orthogonality is achieved by means of Doppler Division Multiple Access (DDMA).

The demo covers antenna array modeling, radar transceiver and waveform design, scenario creation and I/Q data generation, and the signal processing algorithm. It also shows how to accelerate the data generation with MathWorks cloud and parallel computing technologies as well as speeding up the signal processing on GPUs.

Enhancing Quality and Productivity with Shift-Left by Leveraging Pre-Submit Workflow for Continuous Improvement

The shift-left approach emphasizes early testing and quality assurance practices in software development. The demo focuses on the implementation of a well-structured pre-submit workflow to enhance software quality and productivity by integrating robustness in testing and quality assurance processes earlier in the development cycle. Key highlights of the demo include:

• Components of the pre-submit workflow such as static code analysis and code review practices.
• How automated tools and techniques can be seamlessly integrated into the pre-submit phase, including quality gates for continuous integration, code quality metrics, and static analysis.
• How the adoption of a shift-left approach allows developers to get early feedback, proactively address potential defects, improve code maintainability, and enhance collaboration among team members.

Systems Engineering, ASPICE, Requirements-Based Testing and ISO 26262

Performing early verifications on the model level—and later on the code level—to ensure model and code correctness are strategic activities within Model-Based Design. Learn and discuss capabilities such as model checking, coverage measurement, requirements management, and static analysis at the code level to increase the quality of your design. Understand how Model-Based Design with MATLAB^® and Simulink^® supports you in achieving compliance with standards for functional safety (ISO 26262), Automotive SPICE, and guidelines such as MISRA. Additional benefits of implementing a mature issue detection process that identifies issues before going into production are high-quality software and contributing to the principles and metrics of DevOps or SAFe^®.