Microcontroller for Artificial Intelligence

Next generation AURIX™ is a heterogeneous architecture providing new computing abilities for new approaches

As AI-based algorithms require a lot of computing power, Infineon’s AURIX™ TC4x solution is advisable to integrate AI and microcontrollers. Offering high real-time performance and implementing the latest trends in AI modeling, AURIX™ TC4x enables embedded AI for autonomous driving, eMobility, and connectivity, enhancing traditional control and signal processing with data-driven AI. Besides, embedded AI and safety are also highlights at Infineon, which in partnership with the Fraunhofer Institute (IKS), develops safety assurance processes for AI-based systems referring to well-established and new ISO standards: ISO 26262 (safe design for HW & SW), ISO 8800 (safety for AI) and ISO 21448 (safe system specification).

Applications with Embedded AI:

• Traction Inverter
  • Motor control: increased driving distance and battery size reductions with accurate control of torque and speed for energy efficiency. With AI based MPC and MAC, it enables the use o SiC and GaN, as well as the optimization of costs on sensors’ data processing.
  • Virtual sensor: detects aging and temperature of SIC and GaN based power modules, besides increases WLTP driving cycles, allowing cost optimized cooling concepts
• On-board Charger
  • Adaptative control: increased accuracy for in-vehicle power distribution for energy efficiency, optimizing driving distance and battery size reduction
  • Virtual sensor: detects aging and temperature of SIC and GaN based power modules, besides increases WLTP driving cycles, allowing cost optimized cooling concepts

• Battery Management Systems
  • SOX Prediction: AI enables higher accuracy battery state calculation during run-time, leading to higher driving distance or reduced battery size.
  • Fast Charging: AI reduces the charging time with an accurate state of health of the battery.

Edge AI with DEEPCRAFT™

With the Edge AI software brand DEEPCRAFT™, Infineon delivers customers end-to-end Edge AI solutions. By combining with AURIX™ Development Studio, DEEPCRAFT™ Studio provides an end-to-end workflow for creating Edge AI models reducing the time to market. Among the solutions, DEEPCRAFT™ offers improved data collection, support with starting projects with pre-collected and pre-trained models, besides an easier deployment of models to Infineon hardware.

Artificial Intelligence (AI) on TRAVEO™ T2G

Versatile Applications of Artificial Intelligence in TRAVEO™ T2G

TRAVEO™ T2G offers high performance, advanced human-machine interfaces, and robust security for automotive applications, among them: AI applications, including predictive maintenance, intrusion detection, and thermal management, enhancing vehicle performance and security, besides supporting the ARM ecosystem for AI training, development, and deployment via ModusToolbox™.

Versatile Applications of AI:

TRAVEO™ T2G supports various AI applications, enhancing functionalities across different vehicle domains:

• Sensors: AI enables better raindrop detection for rain sensors and superior accuracy in window pinch detection and other Body/Motor control.
• Predictive Maintenance: AI facilitates better prediction of sensor and actuator lifetime, along with health observation of ECUs for earlier maintenance and replacement.
• Gateway & Security: AI improves the accuracy and abnormally decetion for Vehicle Entry systems, as well as intrusion Decetion on IVN (CAN or Ethernet)
• Predictive & Efficiency: AI enables Higher Energy Efficiency as well as Predictive Thermal management

Artificial Intelligence in BCM, Body domain and Zones

• Rain sensor: AI improves rain sensor performance
• Window pinch detection: AI better’s user safety by more accurately predicting window pinch condition
• Networking Security: AI facilitates security by improving detection of attacks
• Predictive Maintenance: AI enables a prediction about lifetime state of local sensors and actuators such as lighting, power elements & drivers
• Predictive Maintenance: Accuracy increase of motor control
• HVAC Loop: AI improves Energy efficiency, better temperature control and in cabin comfort (air mix, flow)
• Thermal Management: AI facilitates better control of cooling & thermal management and better utilization of sensor information

Development flow supported by AI:

• The flow is designed to take your project from data collection to the creation of interfaces. Infineon’s End to End Machine Learning Solution with DEEPCRAFT™ & ModusToolbox ML are under development. The DEEPCRAFT™ makes it easy for you to use machine learning in your edge devices.

PSOC™ Edge the Next Generation of ML Enabled MCUs

PSOC™ Edge series of Arm® Cortex®-M microcontrollers feature high-performance, low-power, secured MCUs with hardware-assisted machine learning (ML) acceleration for next generation applications. The PSOC Edge devices are based on high-performance Cortex-M55, including Helium DSP support paired with Arm Ethos-U55 as well as a low-power Cortex-M33 paired with Infineon’s ultra-low power NNLite hardware accelerator intended to accelerate neural networks in Machine Learning and AI applications.
Support for “always-on” sensing and response makes the devices ideal for advanced IoT devices such as smart home, security, wearables, robotics, and many more. The family of devices come and ample on-chip memories to support most advanced applications PSOC™ Edge consists of dozens of devices that provide scalability for increasingly sophisticated next-generation intelligent systems, while enabling software reuse and portability.
PSOC™ Edge is supported by a rich set of enablement with the industry-recognized ModusToolbox software including integration with the DEEPCRAFT™ Studio AI solution and off-the shelf DEEPCRAFT™ Ready Models.

AI algorithms enable easier design and tuning

Infineon's PSOC™ Automotive microcontrollers (MCUs) integrate artificial intelligence (AI) to enhance human-machine interfaces, smart sensing, and general-purpose applications in vehicles. PSOC™ Automotive 4. AI algorithms within PSOC™ reduce tuning times and support unusual shapes and sizes for trackpads and touchscreens.

Applications of AI on PSOC™ Automotive:

• Sensors for liquid-level sensing, distinguishes between liquid and foam and is insensitive to the size and shape of the liquid container. AI reduces tuning time from weeks to hours for liquid-level sensing.
• Trackpads and touchscreens, support unusual shapes, sizes, and requirements. With AI, algorithm development is reduced to weeks after collecting training data, whereas the traditional approach takes months.

Development flow supported by AI:

The flow is designed to take your project from data collection to the creation of interfaces. Infineon’s End to End Machine Learning Solution with DEEPCRAFT™ & ModusToolbox ML are under development. The DEEPCRAFT™ makes it easy for you to use machine learning in your edge devices.