Kits not bits - fast tracking edge AI development with ease and no risk
Artificial intelligence (AI) and machine learning (ML) are increasingly being deployed at the edge for tasks including machine vision, industrial automation, and environmental control. However, implementing the compute required, particularly for mobile and battery-powered devices, is a challenge.
A system on module (SoM) approach can ease the development of these systems for companies that don’t have the time, budget, or range of specialised skills required to do a ground-up design.
This article looks at how these modules have evolved from development tools, for teaching and prototyping, to enterprise-quality components that can simplify your product development while meeting industrial standards for quality, operating temperature, and so on.
It will also describe how you can use Arduino Pro Kits, exclusive to Avnet, to implement and deploy network enabled products that include sophisticated edge compute, including AI and ML.
To understand the advantages of the module-based design approach, it is worth looking at how Arduino boards have gone from being microcontrollers accessible to anyone, to a reliable enterprise platform for professional developers.
Arduino history
The first Arduino board was designed at the Interaction Design Institute Ivrea (IDII) in Italy about 20 years ago. The goals were to produce a product that was lower cost than the available alternatives (to be affordable by students “It had to be the equivalent of going out to dinner at a pizza place,” according to Massimo Banzi, co-founder of Arduino) and easy to use. The board and software were intended to be used by designers, not only engineers.
In 2010, the iconic Arduino UNO board was launched. The hardware design was open, with all schematics published, allowing others to make compatible products. In particular, add-on boards, known as “shields”, were developed to provide extra peripherals and interfaces. These plug into the headers on the Arduino board and multiple shields can be stacked to combine functions. This makes it easy for users to integrate inputs such as switches and sensors, outputs such as displays and actuators, and network connectivity into their designs.
This was widely used as a prototyping platform for software development, proofs of concept, and small-scale production. A large network of online users shared their ideas, expertise, and designs.
Over the years the range of Arduino boards has grown to meet the demands of different application areas and, in 2020, the Arduino PRO range was launched to provide industrial quality modules with the same low cost and ease of use as the traditional Arduino products. These modules are the ideal solution for implementing advanced IoT edge applications.
- The Portenta family is a series of high-performance industry-rated microcontroller boards. They have two asymmetric cores so they can run low-level real-time tasks, combined with high-level code such as protocol stacks, interpreted languages like MicroPython or JavaScript, and machine learning.
They support a variety of modern wireless communication protocols as well as supporting older wired standards, such as RS-485. This means they can operate in, and provide a bridge between, both legacy and new systems.
With outstanding performance and industry-grade security, the Portenta family enables you to leverage the Arduino ecosystem to deploy powerful AI algorithms and machine learning on the edge.
- Nicla is Arduino’s smallest industrial-oriented board. The Nicla family provides a variety of industrial-grade sensors with processors suitable for low-power AI and machine learning at the edge. A Nicla module can work as an autonomous, battery-powered system but also includes wireless connectivity for easy integration into a wider networked system.
- The MKR range is designed to support IoT projects needing a range of wireless communication interfaces such as Wi-Fi, GSM and LoRa.
- The Opta micro PLC family is targeted at industrial automation and IoT applications where programmable logic controllers (PLCs) have traditionally been used. The PLC IDE for Opta supports 5 standard PLC programming models (ladder, sequential function chart, function block diagram, structured text, and instruction list) for rapid development and porting of PLC applications.
Arduino Pro Kits exclusive to Avnet
To make it even easier to get started with the Arduino Pro range, Avnet provides the Arduino Pro Kits tailored for specific industry requirements. These kits include advanced microcontrollers, sensors, actuators, and software. Rather than a “bag of bits”, you get a complete development solution that combines Arduino Pro products, accessories, tools, and guides into a single product. The combination of powerful hardware and a user-friendly development environment enables you to rapidly design and implement sophisticated industrial systems
The kits support a wide range of application areas using AI from machine vision, to industrial control and smart agriculture. Some examples are described below.
Arduino Pro Edge AI/ML Vision and Speech Kit
The Arduino Pro Edge AI/ML Vision and Speech Kit will enable you to learn and create solutions based on computer vision, speech recognition, and machine learning using the Edge AI tools. This makes it suitable for controlling hands-free or out-of-reach equipment, and to implement accessible systems.
The hardware contains the Portenta H7, Portenta Vision Shield Ethernet, and Nicla Vision. The kit includes software such as the Arduino speech recognition engine, Arduino Cloud access, and machine learning tools, plus technical support.
The Portenta H7 compute module is based on a dual-core microcontroller from STMicroelectronics. Other members of the Portenta family use devices from NXP and Renesas. An incredibly valuable feature of the Arduino middleware is that it ensures interoperability so you can run the same code on all these modules, even when they use different vendor’s silicon.
This kit is suitable for industrial automation such as product inspection and quality control in a production line, predictive maintenance by tracking defect levels, and voice-controlled machines. Other possible applications include building automation and smart cities (for example environmental monitoring and control systems) or patient monitoring in a healthcare environment.
Arduino Pro Industrial Automation and Lighting Kit
The Arduino Pro Industrial Automation and Lighting Kit targets control applications in a micro PLC environment using industrial communication protocols. The hardware includes the Opta RS485, Opta Wi-Fi and a single phase power analyser.
This kit has applications in building management, for example energy monitoring and control, and in industrial control systems with remote monitoring via the cloud.
Smart Agriculture & Irrigation Kit
Smart farming, or precision agriculture, is the use of technologies from industrial control combined with data-driven farm operations to improve sustainability and productivity in agricultural production. This is an increasingly important application for microcontroller technology.
The Smart Agriculture & Irrigation Kit is perfect for managing agriculture applications. It takes advantage of the control capabilities, sensor data analysis and Cloud connectivity of the Arduino Edge Control board to create a low-power system for irrigation control.
For example, you can monitor water levels, program irrigation timers, and monitor water consumption using Wi-Fi connectivity and Arduino Cloud. You can even intelligently schedule irrigation by leveraging weather forecast data available from Arduino Cloud.
Conclusions
If you need to implement AI or ML algorithms at the edge, then the Arduino PRO range can help.
These are industrial-quality modules with the same ease of use as the classic Arduino products. They have sufficient processing power to implement AI at the edge, while being very low-power so they are suitable for remote, mobile, or wearable applications.
To get you started, the Avnet-exclusive Arduino Pro Application Kits provide a complete package of hardware and software for several application areas.
Using a SoM approach simplifies your development process and hence saves you time and money. By removing the need for low-level hardware and software implementation, it lets you focus on delivering innovative products.
Finally, the open architecture of Arduino modules supports your move to a chip-down implementation when a successful product justifies it.
See Arduino Overview
About Author
Harvey Wilson
Harvey Wilson is a Systems Engineer Professional (Smart Industry) for Avnet Silica in the EMEA regio...
silica content library/articles/fast-tracking-edge-ai-development-with-ease
Fast tracking edge AI development with ease and no risk | Avnet Silica
