Signal recognition controller powered with a NeuroMem® neural network to learn and identify splines in time series. The time series are received over Ethernet from a specialised server software running on the MES system. The possibility of using asymmetric encryption schemes is provided for in the data transfer process.
The NeuroMem neural network can be ordered with a capacity of 1024 or 2048 neurons.
The controller comes in an ABS plastic housing with a degree of protection IP20 and it can be mounted on a 32 mm DIN rail.
NT Adaptive is a artificial intelligence (AI) pattern recognition device incorporating a microcontroller ARM of series M4, one or two chips СМ1К, power sources DC/DC, FLASH memory, real time clock, Ethernet port, USB 2.0 port, and LED indication of the status and ongoing processes. Depending on the neural controller model, NT Adaptive can contain 1024 to 2048 neurons. Exchange with the chip CM1K (or chips if there are two), can be carried out either via the serial bus I2C or parallel data bus, 25 lines. Power supply to NT Adaptive: 24 V DC.
NT Adaptive is contained in a housing made of ABS plastic with a degree of protection IP20. The housing of NT Adaptive has the possibility of being mounted on a 35 mm DIN rail.