Smart Home
AI Enhances the Life in the Home
Smart homes make life in the home more safe, convenient, energy efficient, comfortable, fun, enriching, and accessible. Consumers delight in seeing their devices and appliances come to life as emerging AI technology automates, animates, and enlivens them.
Enliven the Smart Home with AI
Today’s smart homes allow users to remotely control and preprogram devices such as surveillance cameras, thermostats, and robotic vacuums. For example, a smart thermostat that knows when no one is home can adjust the temperature to save energy. AI enables these domestic devices to learn from the behavior and activities of home users to provide a better experience through more targeted, more intelligent actions. Whether it’s a security camera recognizing a familiar person versus a stranger or a robotic vacuum navigating obstacles, AI is becoming integral to home life.
Edge AI for a Richer User Experience
Robotic vacuum cleaners, security cameras, smart locks, learning thermostats, intelligent kitchen and laundry room appliances, fitness mirrors, and smart speakers all benefit from the addition of edge AI. In the future, smart devices will become more proactive and take appropriate actions independently. Edge AI enables more device autonomy and decision-making with minimal latency and power consumption, opening the door to a much richer user experience.
Customized AI Processors for the Smart Home
AI processing requirements for smart home applications tend to be less performance-intensive than other markets. However, cost pressures limit the available silicon area for Neural Processing Unit (NPU) engines. Additionally, power consumption must be minimal, especially for battery-powered devices. Smart home devices generally run a single or small group of Neural Networks (NNs). In constrained applications, such as cost-sensitive kitchen appliances, a customized NPU should be specifically designed to use the smallest silicon area (thus, cost) and the lowest power consumption.
Future smart home devices will process ever larger data sets and higher-resolution imagery. As a result, processing at the edge will require highly efficient hardware solutions. Adding Expedera Origin™ E1 or E2 Neural Processing Units (NPUs) to your silicon solution can improve system performance without increasing system costs. With Expedera, a smart home device can run one or more real-time AI inference models simultaneously with minimal power consumption and the smallest area (often 50% or less than other solutions), saving system cost.
An Ideal Architecture for Smart Home
The Origin E1 and E2 neural engine uses Expedera’s unique packet-based architecture, which is far more efficient than common layer-based architectures. The architecture enables parallel execution across multiple layers, achieving better resource utilization and deterministic performance. It also eliminates the need for hardware-specific optimizations, allowing customers to run their trained neural networks unchanged without reducing model accuracy. This innovative approach greatly increases performance while lowering power, area, and latency.
Sustained Low-Power Performance for Always-on Devices
Many smart home devices, such as security systems and doorbell cameras, require continuous sensing and processing.
For always-on applications like these, Expedera offers its LittleNPU E1. The LittleNPU operates at sustained low power (as low as 20mW) while keeping all data within the LittleNPU subsystem, safeguarding user data.
Area-Efficient for Cost-Effective Deployments
One of the hardest challenges in deploying AI is finding a solution that fits the tight budgets of OEMs. Expedera’s Origin NPU requires minimal silicon area, ensuring that AI can be deployed in industrial chips cost-effectively.
Future-Proof with In-Field Updates
With the rapid evolution of AI technologies and networks, any AI deployment must be capable of in-field updates, including new networks. Origin IP is flexible enough to allow the deployment of public, private, and custom neural networks after your device has shipped.
Download our White Papers
