Chip Trio Advances Mobile AI Apps

Artificial Intelligence (AI) is finding inroads to more than manufacturing and industrial automation. Mobile devices and related applications are more than fair game for the technology. For example, vision company NeuroMem Technologies recently introduced NeoCurie, a product line that promises to ease development of mobile AI designs.


The first entry in the product family is NeuroTile, a trio of devices combining features such as low power consumption, advanced learning, and mobile AI capabilities. A mobile AI module targeting wearables, robotics, drones, arms, and vehicles, NeuroTile combines an STM32 microcontroller (MCU) from STmicro, an FPGA from Lattice, and a neuromorphic chip NM500 from NeuroMem/Nepes on a module measuring 35 mm x 26 mm. Using an extra STM32 MCU can provide more software capabilities and Bluetooth functionality.


Sensors 2018 Hits Silicon Valley June 26-28!

Join thousands of engineers this June in San Jose at the sensor industry’s biggest event! With 65+ Technical Sessions, 100+ Leading Speakers, and 300+ Interactive Exhibits, there’s more opportunity than ever to connect with this booming industry and the technologies driving it. See thousands of the newest technologies in action, learn about the latest applications, including AI, Autonomous Vehicles, IoT, and Medical, and develop invaluable partnerships at the only event dedicated to sensors, connectivity, and systems.


NeuroTile uses the pattern recognition and training capability from NeuroMem technology, an AI technology on silicon that allows learning and recognizing pattern in real-time. The Neuromorphic NM500 chip features include: 576 neurons, each one holding up to 256 bytes learned or loaded pattern signature; neither data, nor instructions displacement providing unmatched power efficiency; equivalent to 85 Giga operations/s at 37 MHz and 120 MW; real-time learning capability; novelty detection and high-performance clustering (unsupervised) capability; constant learning and recognition time (8 µs) regardless of network size; and an expansion bus for additional neuromorphic chips (NM500 or CM1K)


According to the company, NeuroTile suits all applications requiring sensors with the need of easy and fast pattern learning and pattern recognition. For more info, contact NeuroMem Technologies.