Microchip Introduces MPLAB® Machine Learning Development Suite

• Blockchain Council
• September 21, 2023

Microchip Technology, a pioneer in embedded control solutions, has ushered in a new era of innovation with the introduction of the MPLAB Machine Learning Development Suite. In a world where machine learning (ML) is rapidly becoming the norm for embedded systems, this integrated workflow offers a streamlined approach to ML model development. Designed to cater to a wide range of microcontrollers (MCUs) and microprocessors (MPUs), this software toolkit empowers engineers to implement ML inference swiftly and efficiently. In this comprehensive article, we delve into the features, benefits, and implications of Microchip’s groundbreaking solution.

Machine learning, often referred to as ML, has emerged as a game-changer for embedded designers seeking to enhance a multitude of products. In today’s landscape, where connectivity and intelligence are paramount, ML at the edge is proving to be a pivotal factor in making products efficient, secure, and power-efficient. This paradigm shift away from cloud-based processing is at the heart of Microchip’s vision for the future.

Rodger Richey, Vice President of Microchip’s Development Systems business unit, emphasizes the importance of on-device ML, stating, “Machine Learning is the new normal for embedded controllers, and utilizing it at the edge allows a product to be efficient, more secure and use less power than systems that rely on cloud communication for processing.”

Microchip’s integrated solution is uniquely tailored to meet the needs of embedded engineers. What sets it apart is its compatibility not only with 32-bit MCUs and MPUs but also with 8- and 16-bit devices, offering unparalleled flexibility for product development.

The MPLAB Machine Learning Development Suite comes to life alongside the MPLAB X Integrated Development Environment (IDE). This dynamic duo allows developers, regardless of their ML programming expertise, to create ML models optimized for Microchip’s microcontrollers and processors. These resource-constrained hardware platforms present unique challenges, and this toolkit addresses them with finesse.

Powered by AutoML, this toolkit eliminates the arduous and time-consuming aspects of model building, including feature extraction, training, validation, and testing. The memory constraints of MCUs and MPUs are carefully considered, ensuring that the resulting models are efficient and small in footprint.

What makes this solution truly groundbreaking is its accessibility. Even those with limited ML programming knowledge can leverage it effectively, saving costs associated with hiring data scientists. However, it doesn’t stop there; experienced ML designers will find the toolkit sophisticated enough to meet their demands.

The MPLAB Machine Learning Development Suite also opens the door to integrating models from TensorFlow Lite into MPLAB Harmony v3 projects, a fully integrated embedded software development framework. This seamless integration accelerates product development and time to market.

Microchip’s commitment to comprehensive on-device machine learning is evident with the MPLAB Machine Learning Development Suite. It joins the ranks of MPLAB Harmony V3 and the VectorBlox accelerator Software Development Kit (SDK), designed to work harmoniously with Microchip’s diverse field-programmable gate array (FPGA) components.

To accommodate a range of users, Microchip offers various licensing options. A free trial version allows evaluation on up to 1GB of data, with 2,500 labels and five hours a month of AutoML CPU time. While this version is restricted to evaluation purposes only, it provides a taste of the capabilities.

For those ready to dive deeper, a standard license, available for $89 a month, offers a substantial boost. It includes 10GB of data, unlimited labels, and 10 hours a month of CPU time, along with the coveted license to deploy models in production.

For the most demanding users, the “pro” license extends the CPU time to a whopping 250 hours a year (20.8 hours a month) and grants the option to output source code instead of relying on a pre-compiled library.

Microchip’s MPLAB Machine Learning Development Suite is a game-changing solution in the world of embedded controllers. Its adaptability to various MCU and MPU architectures, coupled with its user-friendly interface, makes it a powerful tool for both novice and experienced engineers. With this integrated workflow, Microchip has once again demonstrated its commitment to providing cutting-edge solutions that redefine the possibilities of embedded systems.