Spotlight on SAMA7G54 and Edge Impulse: Driving AI Innovation at the Edge

In the drive to accelerate machine learning (ML) and artificial intelligence (AI) adoption at the edge, Microchip and Edge Impulse have partnered to fully integrate the SAMA7G54 32-bit microprocessor unit (MPU) into the @Edge Impulse platform. Software engineers at all levels can take advantage of?Edge Impulse’s easy-to-use toolset for training, evaluating and deploying ML models on embedded targets, greatly simplifying the model development, selection and optimization process. Edge Impulse’s toolset supports a wide range of applications like computer vision at the edge, audio and signal processing, and more.

Next, is an overview of the AI integration process including data collection and preprocessing, model training, real-time model testing and over-the-air updates.

Data Collection and Preprocessing?

The journey begins with data collection and preprocessing - a vital step prior to constructing your model architecture, which encompasses the input, output, processing modules and learning blocks. The SAMA7G54’s support for various sensors and peripherals enables data to be collected directly from the environment and then sent to the Edge Impulse platform for preprocessing. Within Edge Impulse Studio, developers can easily upload and label this data, and use tools to augment and preprocess it - both essential steps for creating robust ML models.

Dig deeper:

?? Microchip guide: Getting Started with Microprocessors (MPUs)

?? Edge Impulse Documentation: Microchip SAMA7G54

Model Training

With data in place, the workflow moves into model training, with Edge Impulse’s automated pipelines taking center stage. Edge Impulse’s automated training pipelines offer a variety of algorithms and architectures that are optimized for use with edge devices. The platform also provides tools to optimize models so that the edge device can better leverage the hardware capabilities of the SAMA7G54, such as ensuring efficient use of memory, processing power and other resources.

Dig deeper:

?? Microchip blog: AI/ML at the Edge for 32-bit Microprocessors, Using Edge Impulse

?? Press release: "Edge Impulse Integrates Microchip’s SAMA7G54 Microprocessor Into Its Platform to Easily Train Machine Learning Models"

Real-Time Model Testing?

Once the model is trained, the SAMA7G54 MPU can be used to perform real-time testing into how the model performs on the actual hardware. Detailed performance metrics–including accuracy, latency and resource utilization–are available to help developers fine-tune their models for optimal performance on the SAMA7G54 MPU. Deploying the model to the device and running inference directly provides immediate feedback and the crucial insights necessary to accelerate this optimization process.

Streamlined Deployment

Deployment marks the final stage of the workflow, where the combined power and simplicity of the SAMA7G54 and Edge Impulse again comes into play. Edge Impulse generates optimized code and binaries that can be flashed directly onto the SAMA7G54 MPU. This simplicity is further enhanced with support for over-the-air (OTA) updates, which allows developers to push updates and improvements to devices in the field, without requiring physical access.

Dig deeper:

?? "Edge Impulse Partners with Microchip to Offer Support for On-Device AI/ML on the SAMA7G54" from @Hackster.io

With the deployment process simplified and automated, developers can now move from development to real-world application with ease.

Requirements for Empowering AI at the Edge

The ability to streamline ML model creation and facilitate real-time model testing and seamless model deployment is a critical development for an industry that has until now struggled with these processes. The combination of Microchip’s MPUs with the Edge Impulse platform solves these challenges, empowering AI innovations through the following essential capabilities:

● Low-Power Operation With High Performance and Versatility: The SAMA7G54 MPU is designed for low-power operation in AI applications. It is ideal for battery-powered AI applications like wearables, IoT devices and remote sensors. With its 32-bit architecture and advanced processing capabilities, the MCU can handle complex AI models and real-time data processing at the edge. The MPU also supports a wide range of interfaces and peripherals, giving it the versatility to be used in diverse applications, from industrial automation to consumer electronics.

● Scalability: Edge Impulse’s platform is designed to scale, allowing developers to manage multiple devices and models efficiently. This is particularly useful for large-scale deployments such as smart cities and agriculture, among other applications.

● Community Support: Both Microchip and Edge Impulse support their active communities with extensive documentation, ensuring that developers have the support and resources they need at every stage of their AI projects.

Accelerate Your AI Development with SAMA7G54 and Edge Impulse

By integrating the SAMA7G54 MPU with Edge Impulse, Microchip and Edge Impulse are streamlining the AI model development lifecycle from data collection to deployment. This is critical not only for accelerating time-to-market for AI-powered products, but also empowering developers to drive meaningful innovation at the edge as they unlock new possibilities for AI across a wide range of applications and industries.

Stay informed about the latest in edge AI, ML and other advancements and insights spanning Microchip’s broad range of solutions by subscribing to the Microchip Insider LinkedIn Newsletter. Join us to receive exclusive insights, updates and resources that help drive your projects forward.