A New Approach to Edge AI

As Edge AI moves closer to real devices, manufacturers need intelligence that is lightweight, adaptive, and able to learn directly from real-world data.
PicoAI, developed by Boon Logic, brings self-learning capabilities to low-power embedded systems — and becomes even more powerful when combined with MicroEJ VEE Energy.

What Is PicoAI?

PicoAI is an unsupervised Edge AI engine designed for microcontrollers.
It learns directly from live data, without cloud training or labeled datasets.

With PicoAI, embedded devices gain the ability to:

• Learn patterns locally
• Adapt over time
• Detect anomalies and abnormalities
• React to data quality degradation

This enables responsive, autonomous behavior on cost-effective hardware.
 

Watch the PicoAI Demo on NXP i.MX RT1170

In this demonstration, PicoAI trains in real time using household metrology data on an NXP i.MX RT1170 running MicroEJ VEE Energy.
The video also shows live visualization through a Grafana dashboard and how the engine reacts when data becomes corrupted.
 

What the Engine Can Detect

Using its self-learning capabilities, PicoAI extracts structure from real-time data streams.
In practical demonstrations, the engine can:

• Detect on/off events
• Estimate activity contributions (e.g., solar variations)
• Distinguish evolving consumption patterns
• Identify data corruption through a built-in quality score

These behaviors occur automatically, without predefined models.

Why PicoAI Stands Out in Edge AI

True On-Device Learning
PicoAI trains directly on the device — no cloud resources, retraining cycles, or large datasets required.

Ultra-Lightweight for MCUs
The engine operates on microcontrollers like the NXP i.MX RT1170, making Edge AI accessible even for cost-optimized products.

Adaptive and Explainable
The engine provides immediate feedback when data becomes unreliable, enabling safe decision-making in dynamic environments.

Enhanced by MicroEJ VEE Energy
With MicroEJ’s embedded container technology, manufacturers benefit from:

• A secure runtime
• A fast, modern UI
• Cloud connectivity
• Easy portability across multiple MCU/MPU platforms

Together, PicoAI + MicroEJ provide a practical roadmap to scalable Edge AI products.

A Practical Path to Edge AI at Scale

PicoAI and MicroEJ open the door to deploying adaptive, self-learning intelligence directly within embedded devices.
This combined approach supports reliable, efficient, and future-proof Edge AI implementations across a variety of embedded domains.

Learn More

▷ PicoAI technical overview
▷ MicroEJ press release on unsupervised Edge AI