Megatrend electrification: smart system integration, AI and cybersecurity as trailblazers

Electrification is gaining momentum in many industries and is driving independence from fossil fuel-based technologies. With the use of power electronics, batteries and electrical machines of all sizes, it has become commonplace to use a growing number of embedded systems for control and regulation. How can engineers design such complex systems, iterate quickly and validate their designs along the way? For many developers in industries like renewable energy, mechatronics and transportation, the answer is model-based design.

Frank Graeber

Progressive electrification presents two challenges: the development of individual components and the optimization of the overall system. In traditional component development, requirements and interfaces are clearly defined. The aim is to implement these components quickly and efficiently, often using model-based design. However, the true complexity only unfolds in the system-wide optimization, which includes not only individual vehicle components but also the interaction of the vehicle, charging infrastructure and energy generation.

The development of systems and components can no longer be viewed separately; instead, they must be interlinked. Electrification therefore crosses domain boundaries and different technologies. Looking at the system as a whole helps to define requirements and develop a deeper understanding of component boundaries. A comprehensive system design makes it possible to integrate the different requirements of manufacturers and hardware platforms and to make a manufacturer-independent hardware selection.

Link system requirements to architecture models to establish requirements traceability and perform requirement coverage analysis or impact analysis.

Hardware independence

In addition to efficient system modeling, model-based design enables hardware-independent algorithm development: algorithms implemented on an embedded system can be developed and verified independently of the hardware platform. This enables the integration of controllers as components that don’t yet have to be implemented as specific hardware in the model. Only in the second step is the algorithm implemented on a specific hardware platform with the help of hardware support packages. These support automated code generation and virtual hardware-in-the-loop simulation.

This approach is already widely used in the automotive industry. The move toward battery electric vehicles (BEV) and autonomous driving (AD) is leading to a growing need for new electrical/electronic (E/E) architectures, increased computing power and increasingly complex algorithms. In addition, recent disruptions in the semiconductor automotive supply chain have highlighted the importance of semiconductors and the need for cross-platform flexibility.

AI in electrification

In many respects, the use of artificial intelligence (AI) is setting new standards for the further development of electrification. For example, AI methods that enable the creation of models for physical components or systems significantly accelerate the simulation process while mapping the essential behavior of the system. The integration of AI models into validation and testing also offers advantages. The development of virtual sensor systems using AI is another innovative field. These provide critical signals for the control of electrical systems without recurring material costs or maintenance.

AI tools also enable the measurement of variables that are difficult to physically measure, such as the state of charge or the state of health within a battery management system (BMS). The creation of physics-based models and the generation of synthetic data for the training of such sensors are crucial. These models can be directly integrated into system models for validation.

AI-based approaches are also used for the development of control algorithms in non-linear systems, whereby models can be trained against simulated environments and integrated into simulation-based verifications. In the field of energy forecasting, AI applications help predict the power supply, demand and pricing more precisely, which leads to a reduction in uncertainties in electricity grid operations.

Cybersecurity in system development

The increasing networking of electrical systems also requires comprehensive consideration of cybersecurity to prevent cyberattacks and ensure the security of the infrastructure. Innovative approaches and the integration of security aspects in all phases of system development are required here.

Cybersecurity must be considered at every stage of system development, starting with system simulation and requirements engineering. Fortunately, the importance of cybersecurity is already being underpinned by the first set of rules, standards and regulations. It’s to be expected that legal regulations in this area will continue to increase. During implementation, the vulnerabilities identified in the system design must be secured by secure programming practices. Programming rules and robustness play a decisive role there.

Demos at Embedded World

Would you like to see Mathworks services and products live? At Embedded World booth 4-110, you can get an insight into the following demos: - Integrated model-based control workflow including HIL tests. This demo presents a fully networked model-based engineering workflow for the development of digital controls for motors and drives. - Next-generation BMS for fleet and cloud integration. This demo extends a ‘classic’ battery management system to adapt it to the software-defined vehicle trends. - Cross-platform workflow for motor controllers. This example demonstrates an embedded targeting workflow for motor controls that emphasizes cross-platform flexibility, based on the Motor Control Blockset. - Accelerate your Autosar development with Mathworks Consulting. Mathworks Consulting provides expert support for Autosar development, optimizing the process and accelerating software development. - Polyspace static analysis and dynamic testing of C/C++ code in CI. This demo shows the possibilities of Polyspace for static code analysis and dynamic unit testing for C/C++ code. - Ground plane segmentation of lidar data on FPGA. This demo shows the segmentation of an organized 3D lidar point cloud into ground and non-ground portions on an FPGA. - Implementation of AI-based motor control on Infineon Aurix TC4x. This demo shows a verification and implementation workflow to implement an AI-assisted field-oriented controller on the Infineon Aurix TC4x heterogeneous multicore microcontroller platform.

Solution strategies for an electrified future

On the way to a revolutionary, electrified future, it’s all these integrated solutions that will enable a fast, efficient and safe implementation. Model-based design and AI are becoming pillars of development that not only overcome the boundaries between different technologies and manufacturers but also ensure the sustainable and safe operation of electrical systems. With these tools and methods, engineers and developers are ideally equipped to drive electrification forward and set the course for a more environmentally friendly and technologically intelligent world.