Programs on the device
BASIC programs are tokenized and stored in EEPROM. Execution context survives cooperative slices, leaving main-loop time for communications and I/O.
03 / Super AI PLC
Super AI PLC
Control with edge learning
Programmable control with on-device statistical learning: measured signals become learned baselines, deviation indicators and conditions usable by the control program.
Product images and views
Select an image to view it at full size.
Architecture and technology
Learning a signal baseline
ANOM updates a signal mean and variation using EWMA. Warmup and consecutive-deviation confirmation turn a departure from the baseline into a program variable.
Local model lifecycle
The learned baseline can be saved, loaded and frozen. Initial learning can be separated from monitoring, with relearning when operating conditions change.
Connecting to wider analytics
MCU analytics detects changes in signal behavior. Vision, language models and heavier analytics are integrated through server or AI services, with execution location defined in the project design.
From input to outcome, step by step
Select a stage to see its role in the system.
01
Signal
A measured value enters the program.
Models and configurations
Reference MCU implementation of control and anomaly detection
System integration with analytics and AI services
Applications and selection
Signal deviation detectionEquipment behavior monitoringAdaptive control logic
Selecting for your project
Specify firmware, signal type, learning period and alarm response; evaluate detection performance against project data.
Discuss project requirements