TinyML Talks Ukraine - Practical TinyML for Industrial Safety and Personnel Performance Monitoring

This talk presents Leantegra’s transition from custom MCU-side machine-learning code to a TensorFlow Lite-based TinyML stack for industrial RTLS devices. It covers IMU-based motion recognition, end-to-end deployment, model validation, hardware constraints, and industrial safety use cases.

Syllabus

Introduction
Strategic partners
TinyML Summit
TinyML Talks
About Ukraine
Speaker
About our company
Case study DTEK
Hardware Software
Hardware Examples
Software Examples
RD Activities
IMU Theory
Additional Details
Other Use Cases
Motion Recognition
Demo
Questions
Battery life
Common filter
Endtoend deployment
Data labeling
Model validation
Hardware chipset
Device status monitoring
Certifications
Firmware upgrades
Local data storage
Collision avoidance
MQTT broker
Data collection
Poll
Model size
Future releases
Collision avoidance beacons
Summary