Autonomous and automated driving is a megatrend in the automotive industry, along with electrification. AI plays a critical role in this trend, enabling vehicles to detect pedestrians, analyze driver behavior, recognize traffic signs, and control trajectories, among many other use cases. A key enabler of this is Edge AI, as autonomous and automated driving is highly dependent on the need for AI systems with machine learning capabilities and processors that can handle large amounts of data in parallel, safely, secured, and in real time. To address this challenge, Imagimob, an Infineon Technologies AG company, has enhanced its automotive machine learning portfolio by integrating machine learning capabilities into Infineon’s Automotive ASIL-D complaint MCUs like AURIX TC3x and AURIX TC4x.
“The integration of secured and dependable AI capabilities into microcontroller families is crucial for advancing autonomous driving applications in the automotive industry,” said Thomas Boehm, Senior Vice President Microcontroller at Infineon. “We are proud that our AURIX microcontrollers are now supported by Imagimob Studio, making them accessible to developers worldwide. This highlights our role as a leading innovator in the industry.”
“With the integration of AURIX into our Imagimob Studio, we are bringing full machine learning (ML) compatibility and capabilities to the automotive sector,” said Alexander Samuelsson, CTO of Imagimob. “This means that all the use cases we support with our platform are now also available for Infineon’s AURIX microcontrollers.”
With Imagimob Studio, developers can now create robust ML models for the Edge and deploy them onto Infineon’s proven AURIX MCUs. The process starts with creating machine learning models in Imagimob Studio. Once the AI model is complete, users can select to deploy on the MCUs directly within the platform. They are then guided through steps on how to deploy the code seamlessly, simplifying the implementation of machine learning on MCUs and enabling the creation of sophisticated ML models. In addition, Imagimob Studio offers a sample project for siren detection, demonstrating model creation and deployment. By using the code example, users can also learn how to create acoustic models with AURIX MCUs and a microphone shield. Furthermore, Imagimob has developed new regression models that can be used to calculate remaining battery power, health status, and usage time.
Advanced AI use cases with AURIX TC4x
The AURIX TC4x scalable MCU family offers a seamless upgrade path from the AURIX TC3x family of ASIL-D compliant automotive MCUs. This enhanced performance is powered by the next-generation TriCore™ 1.8. In addition, the AURIX TC4x features a scalable accelerator suite that includes a parallel processing unit (PPU) and multiple intelligent accelerators to support cost-effective AI integration. For the AURIX TC4x family, these advancements translate into enhanced machine learning capabilities, enabling developers to deploy multiple models simultaneously or more complex ones. For instance, while the AURIX TC3x can handle basic siren detection, the AURIX TC4x enables both siren detection and voice interaction simultaneously.
