Issue link: https://resources.mouser.com/i/1538449
Microprocessor units (MPUs) provide much higher computational power. They use full operating systems such as Linux, allowing them to be far more flexible and programmed for a wide array of tasks. However, this means that MPUs have higher power requirements, and their reliance on external memory and peripherals makes integration more complex. Designers of factory automation must understand the appropriate technology to employ. When considering the need for AI, they must select AI systems that are suitable for their intended task. Many factory applications do not need high levels of AI. In the industrial world, complex language-based AI models are not needed. Instead, compact and efficient AI installed at the edge is much more suitable for roles in the factory, such as data analysis, anomaly detection, and image-based tasks. These AI models are lightweight and optimized to be more energy efficient and less resource hungry. They are small enough that they can be executed on the latest generation of MCUs, enabling the deployment of AI at the edge. Installing the technology close to the point of use means lower latency—the delay caused by the need to transmit data to a remote processor and return the results. Low latency is especially critical on fast- moving production lines or in situations where an instant response enhances safety. The short transmission distances that are enabled by edge AI systems improve energy efficiency, thus removing the need for high-speed data links with the cloud. Edge AI also has a positive impact on overall reliability, as operations can continue even when network connectivity is limited. C h a p t e r 1 | I n t e l l i g e n c e a n d A u t o n o my o n t h e Fa c t o r y F l o o r Markus Mayr Director of Marketing, MCU/MPU, Americas, STMicroelectronics In the past, computer vision algorithms would need an MPU. The STM32N6 is the first microcontroller to deliver this performance at a lower cost point, along with a lower system cost, because fewer external components are needed." 8 Autonomy Meets Intelligence: Enabling the Future of Factory Automation