NeuroShield is a board with the Arduino format and 576 NeuroMem neurons ready to learn pattern vectors or recognize them.
Access to the NM500 is made through a simple SPI protocol and simple functions such as Learn, BestMatch, Classify, Save Knowledge, etc.The network can be expanded by stacking some NeuroBrick modules.
The pattern vectors can derive from any type of signals, whether transmitted by the on-board Accelerometer/Gyroscope or from other Arduino shields with sensors including IMU, audio,environmental sensors, bio-signal, and video.
Recognition results are available within microseconds after the broacast of a new pattern. They can report a simple Known/Unknown status, the category of the closest match or a detailed classification of K closest matches.
Learning can be done in real-time or off-line on data previously collected to the SD card.
- NM500 chip (576 neurons); network is expandable to 4032 neurons by stacking up to 3 NeuroBrick modules
- SD card slot to save and restore the knowledge built by the neurons, collect data, etc
- 6-axis Motion Sensor (InvenSense PMU6050)
- 8 LEDs indicators
- Arduino connectors
- USB_C connector
- Power source
- Arduino/Mbed board
- micro USB
- Read More>>
NeuroShield is a product of nepes (Korea)