Tag - RK3566

aihub z3 vs z5 edge computing box selection

AIHub-Z3 vs AIHub-Z5 Edge Computing Box Selection

The difference between AIHub-Z3 and AIHub-Z5 is not only RK3566 vs RK3588. It includes NPU capacity, peripheral expansion, operating environment, display output, and long-term maintenance boundaries.

Blog, IoT Protocols and Interoperability , June 8, 2026 , AIHub, AIoT, Edge AI, Edge Computing, Industrial Vision, RK3566, RK3588

edge ai observability remote diagnostics

Why Edge AI Fails More Often on Observability Than on Model Accuracy

Edge AI deployments rarely fail first on model accuracy. They fail when teams cannot see input health, inference health, version context, or diagnostic evidence. This article explains why observability should be designed as a core Edge AI capability from ESP32-class devices to Linux edge boxes.

Blog, IoT Protocols and Interoperability , April 9, 2026 , device health, Device Logs, Edge AI, Edge Gateway, ESP32, Incident Triage, Observability, Operations, Remote Diagnostics, RK3566

Edge AI Device OTA: How to Do Staged Rollouts and Rollbacks

The hard part of Edge AI OTA is not pushing a new package. It is designing staged rollout, rollback, and remote recovery for devices whose firmware, model, and configuration evolve together. This article explains how to do it from ESP32 to RK3566.

Blog, Embedded System Development , March 26, 2026 , AI Deployment, Device Operations, Edge AI, ESP32, Fleet Management, OTA, Remote Recovery, RK3566, Rollback, Staged Rollout

YOLOv8 INT8 Performance on RK3566（Detection）

RK3566 Can Run YOLOv8 INT8 — But Only Within These Limits

Real-time YOLOv8 INT8 on RK3566 is possible—if you manage execution paths, quantization effects, and model limits.

Blog, Edge Computing and Data Analytics , January 23, 2026 , Edge AI, Embedded Vision, INT8 quantization, NPU Optimization, Object Detection, RK3566, RKNN Toolkit, YOLOv8

Migrating MediaPipe Gesture Recognition to RKNN: From .task to RK3566 Deployment

A step-by-step guide to convert MediaPipe Gesture Recognition models to RKNN and running real-time inference on RK3566 NPU, with code and troubleshooting tips.

AI and Machine Learning, Blog , December 2, 2025 , Computer Vision, Edge AI, embedded AI, Gesture Recognition, Landmark Detection, MediaPipe, Model Quantization, RK3566, RKNN, Rockchip

How to Deploy YOLOv8 on RK3566: Build Efficient Edge AI Inference from Scratch

Learn how to deploy YOLOv8 on RK3566 step by step. This complete edge inference tutorial covers model conversion, optimization, and real-time detection setup for embedded AI systems.

Blog, Edge Computing and Data Analytics , October 14, 2025 , Computer Vision, Deep Learning, Edge AI, Inference, Model Deployment, NPU, RK3566, RKNN Toolkit, YOLOv8