Machine Learning Embedded Processing
Silicon Labs
Bluetooth Channel Sounding Demo
Shipping/Box Drop/Logistics AI/ML demo
Magic Wand AI/ML – using Bluetooth
Bluetooth Channel Sounding Demo
Channel Sounding, previously referred to as High Accuracy Distance Measurement (HADM), uses Phase-Based Ranging (PBR), Round Trip Time (RTT), or both to accurately measure the distance between two Bluetooth Low Energy connected devices.
- It enables connection-oriented 2-way ranging.
- Supports up to four antenna paths between devices – minimizes multipath effects and enhances accuracy.
- Offers enhanced built-in security features to mitigate the risks of man-in-the-middle or relay attacks.
AI Parcel Tracking xG24 Demo
This tracking solution is a collaboration between Silicon Labs and NeutonAI. Designed to identify seven unique events during the package delivery process, this revolutionary solution combines #siliconlabs ‘ EFR32MG24 Wireless SoC with Neuton.AI ‘s compact neural networks. With a footprint of under 3 kilobytes and an inference time of less than 1 millisecond, this integration provides a comprehensive business solution.
The 7 states of parcel delivery:
- Parcel IDLE (no movements) / IDLE in wrong orientation
- Parcel Shaking
- Parcel Shocked (Damaged)
- Parcel Free Fall (Works if the parcel was dropped from a height of about 1 meter)
- Parcel Transported by Courier / Transported by Courier in wrong orientation
- Parcel Transported by Car / Transported by Car in wrong orientation
- Unknown state
Magic Wand AI/ML – using Bluetooth
This Magic Wand detects gestures drawn in the air, using AI/ML, to control a light over Bluetooth.
Wireless Factory Automation
Microchip
This demonstration shows off Microchip’s range of different wireless MCUs and their on the edge AI and ML capabilities.
Why ML? ML is a set of algorithmic methods that discovers patterns from seemingly unrelated data, providing you with important information to facilitate decision making.
Why on Edge? ML on edge makes the system power efficient, fast and secure. User privacy is at the forefront because personal data never leaves your device. ML on edge also saves cloud resources and compute power in storing and maintaining data pipelines.
Why Microchip? We offer 8-, 16- and 32-bit microcontrollers (MCUs), microprocessors (MPUs) and Field-Programmable Gate Arrays (FPGAs). With a simple ML design process that can bring an ML engine to each of these systems quickly and efficiently, we offer solutions for a wide spectrum of users such as embedded systems engineers and data scientists. Our AutoML-powered design process automates the steps to build the ML model and will go through multiple iterations until a satisfactory model is identified.