Home is where AI is

Very soon most homes will have Siri, Alexa, Google Home or similar. Many already have all three. The acceptance of such sophisticated AI systems as an everyday, normal addition to the living room says much about the human condition to imagine, conceptualize, innovate, experiment with and finally to accept and become familiar with fundamentally new ideas and new ways of living.

We are just learning what we can achieve with AI at the edge. Currently we ask: ‘Siri, what’s the time?’; ‘Alexa, what’s the weather like?’; ‘Google, please set the timer for 30 minutes’ But humans constantly push the envelope and AI systems are capable of much more. So much more.

Is that a cat moving about inside, or is it an intruder? Face detection can determine which. Who is that ringing my doorbell? Facial recognition will ensure you let friends in while sending cold-callers or burglars next door where your neighbors haven’t embraced the technology.

Smart Home AI applications are limited only by the imagination. Anywhere you can place a sensor – heat, light, movement, camera, etc– and link it to network, there is a place for AI, making our lives easier, more convenient, safer, more fun.

And it’s never been easier, fast or cheaper to develop ideas into real products. Take the Face Detection example. Using ultra-low power Lattice FPGA solutions it’s possible to embed AI-powered Image Detection in an image sensor environment in just weeks, leveraging an advanced neural net engine that is innately more accurate than traditional algorithms, and so detect and confirm any specific close-proximity item – face, animal, object – in real-time.

Because all the processing takes place locally in the device, only frames containing a confirmed item are streamed to the cloud, lowering cost. And at 5 frames per second, the Lattice processing device, iCE40 UltraPlus consumes only 800 microwatts of power.

Lattice has streamlined the path for system designers to get started with the Lattice iCE40 UltraPlus™ Mobile Development Platform (MDP) hardware plus the Lattice sensAI stack, the world’s smallest deep neural network engine. With free soft IP cores (BNN accelerators, CNN accelerators), software tools, reference designs, demos, and customized design services, sensAI accelerates integration of on-device sensor data processing and analytics in edge devices.