Edge AI specialist Syntiant is collaborating with Danish industrial ceramic bearings manufacturer CeramicSpeed to develop ultra-low-power autonomous sensors for Industrial IoT applications using the Syntiant NDP120 Neural Decision Processor. The collaboration Aimed at Reducing Costs through Predictive Maintenance in Industrial IoT Applications.

Code named the Bearing Brain Project, CeramicSpeed plans to offer remote monitoring of industrial machinery with NDP120-embedded wireless sensors that will use AI to learn, detect and respond to operational anomalies at sub-milliwatt power consumption.

Built using the Syntiant Core 2 programmable deep learning architecture, the NDP120 is designed to natively run multiple deep neural networks on a variety of architectures, such as Convolutional Neural Networks (CNNs), Recurrent Neural Network (RNNs) and fully connected networks. With support of up to 7 million parameters, the NDP120 processor provides the neural capability for accurate inference on challenging tasks such as multi-sensor processing and can run multiple applications simultaneously at under 1mW (milliwatt).

“The NDP120 stands out as the most flexible, neural co-processor available for battery-powered devices on the market today,” said Morten Opprud, tech lead of the Bearing Brain project at CeramicSpeed. “Syntiant’s deep learning technology enables us to easily develop highly accurate, low-power sensor devices for cloud-free edge processing that will yield meaningful predictive maintenance data in real time, while also allowing a battery lifetime of several years. The ability to easily program the NDP120 along with the capability to connect either microphones directly via PDM or accelerometers through the SPI in the built-in cortex M0 CPU gives us great flexibility during our initial deployment.”

“Until now, predictive maintenance and condition-based monitoring usually has been done in the cloud,” said Kurt Busch, CEO of Syntiant. “Our NDP120 is ideal for industrial condition-based monitoring applications like the Bearing Brain project, enabling CeramicSpeed to move prediction and forecasting down to the battery-powered sensor device itself to reduce or eliminate unforeseen maintenance costs. Our technology can continuously monitor sounds, vibrations and even temperature with minimal drain on power consumption, extending battery life by months or years, while improving performance, productivity and efficiency across a wide range of manufacturing applications.”