NeuroTile

Smallest learning and recognition engine Pre-Order Now

Real-Time Machine Learning (RTML), battery powered

Cognitive Sensing & Cognitive Storage (patent pending)

2 NM500 CSP (300 GigaOps equivalent at < 100 Milliwatts)

Applications

  • Speech recognition
  • Abnormal sound detection
  • Gesture recognition
  • Motion monitoring for human, animal, vehicles(robots, drones, etc…)
  • Vibration monitoring (anomaly detections and real incremental learning)
  • A combination of the above and more…

Architecture

  • STM32 ARM CORTEX M4 RISC (low power)
  • Real Time ML AI unit: 2 NM500 (1152 neurons )
  • Micro SD Card on the back:  up to 512 GBytes
  • LBT: ST Blue NRG1
  • 3D accelero., 3D Gyro, 3D magnetometer, barometer, microphone (MEMS)
  • Arduino compatible IDE and additional libraries (C, Python)
  • Software: NeuroTile Lib by GV (Arduino IDE  and Eclipse)  and NeuroMem Knowledge Builder

Download datasheet