Zypher EdgeAI Demo
An embedded gesture-recognition demo running on an emulated Arm Cortex-M.
ZephyrTensorFlow Lite MicroQEMUCMSIS-NN
Context
This edge AI demo runs gesture recognition on an emulated Arm Cortex-M3 with Zephyr RTOS and TensorFlow Lite Micro.
Approach
The system uses a quantized model, RTOS threads, structured UART output, runtime diagnostics and a QEMU-based test workflow. It recognizes four gesture classes using accelerometer input.
Scope and limitations
The documented QEMU demo uses mock sensor data; it is not evidence of physical-device or field performance.