MikroE adds AI to Necto development tool – video

MikroE adds AI to Necto development tool – video

Technology News |
By Nick Flaherty

MikroElektronika has developed rule based AI, natural language processing and machine learning to increase the efficiency of its development environment as the first step to automatic code generation.

MikroE produces a new development board every day, and managing the documentation and support for over 1700 separate microcontroller and peripheral boards is a key challenge, CEO Neb Matic tells eeNews Europe.

Version 3.0 of the Necto Studio Integrated Development Environment (IDE), launched at Embedded World (EW2023) in Germany this week, now includes an AI Assistant that uses natural language processing (NLP) for a question and answer help service. This uses data from Necto help files, mikroSDK help and Click board help, as well as data from microcontroller datasheets such as registers information, tables and more to answer questions that may be vague or misspelt.

It does this by taking the PDFs from suppliers, breaks them down into up to 10,000 individual, interlinked files that are indexed with the other material to deliver the requested information from the NLP engine.

All the technology was developed in-house, says Matic. “Our goal is to save time, perhaps five minutes per question,” he said. “We have been working on this for over two and a half years.”

Users can ask anything in natural language form, and the AI Assistant will provide a list of potential answers that best match the question. Users can also select a function in their code or Library manager, and with the F1 button receive the best answer. It will also continuously learn from user’s questions, says Matic.

“It’s a bit like Microsoft’s Chatbot,” he said. “However, the AI Assistant is dedicated to the embedded electronics industry, and if you are using MIKROE’s Clicks, SiBRAIN, DISCON or other products you can be sure that answers will be correct,” in dig at the inaccurate answers from natural language tools such as ChatGPT.

This is the first step to AI-driven automatic code generation which the company is already using internally.

Necto Studio 3.0 also adds is an application output panel that lets users view debug-related information and trace information directly from Necto Studio. Used with the Planet Debug remote board tool this provides the trace logs from remote Click boards so that writing debugging or tracing information to the application output window in the Studio tool is now possible. Users can now choose between UART or application output window as an output channel for logging messages from a project using a simple PrintF command.

A custom board tool which allows users to add their custom-designed boards within a mikroSDK and Necto Studio set-up. The wizard-like tool guides users through the steps needed to generate and install all the required code. Custom boards can be shared, or even made public and included in the next official version of mikroSDK.


Linked Articles