The PSoC 6 AI Evaluation Kit provides a range of tools that are required to build intelligent consumer, smart home and IoT applications. It executes inferencing next to the sensor data source, providing benefits such as enhanced real-time performance and power efficiency compared to cloud-centric solution architectures.
Measuring 35 mm x 45 mm it is both small and affordable and its broad range of integrated sensors and connectivity make it suitable for in-field data collection, rapid prototyping, model evaluation and solution creation.
The PSoC 6 AI Evaluation Kit also leverages Infineon’s advanced microcontroller (MCU), sensor and connectivity portfolios and powerful software development environments, including the ModusToolbox and the company’s Imagimob Studio offering for custom ML models, as well as off-the-shelf Ready Models.
“Infineon is helping drive the evolution of Edge AI, which requires the real-time compute capability and power-efficiency that distinguish our hardware portfolio, along with development tools that simplify and speed application development,” said Steve Tateosian, SVP of IoT, industrial and consumer microcontrollers, Infineon Technologies. “With our newest evaluation kit, we are offering access to a complete, easy-to-use ecosystem of Infineon hardware and software that will accelerate development and deployment of intelligent solutions in consumer, smart home, commercial and industrial IoT applications.”
The Evaluation Kit, based on the Infineon PSoC 6 MCU, supports development using the company’s wide portfolio of XENSIV products for automotive, industrial and consumer applications, as well as connectivity including Wi-Fi, Bluetooth, and Bluetooth Low-Energy (BLE) solutions.
With Imagimob Studio, developers can move to production much faster as the platform is free to use (users only pay for new models once they launch in their product) making it easy to build high quality AI models from scratch or optimise existing models.
Imagimob Ready Models also make AI models available to companies without the time, cost or machine learning know-how needed to create custom models.
