Ready-to-customize Edge AI solutions
Run these Edge AI solutions on Cortex M-series MCUs.
Reality AI offers a set of pre-built Edge AI solution frameworks for solving a number of different product challenges. Use them as-is, tweak them to meet your specific requirements, or work with us to create something completely unique to your product.
Looking for more details on a specific industry?
Edge AI solutions for Automotive
Emergency Vehicle Detection
Detect sirens from emergency vehicles and localize them. Identify the direction of sounds as originating in front of, behind, to right or to left.
Our Edge AI solutions can handle reflections and refractions caused by buildings and other obstacles. You can customize to specific microphones, different geographies (siren patterns), or improve them with additional detection targets.
Terrain Detection and Traction Classification
Identify specific terrain under vehicle according to factors affecting ride quality and traction. Identify wet, snow, ice, sand, gravel. Classify pavement by road type.
Applications include ride management, autonomous operation, and smart-city applications. Uses exterior microphone mounted to vehicle and/or accelerometer attached to suspension, depending on specific detection/classification targets.
Condition Monitoring and Wear Detection
Identify the condition and state-of-wear of automotive components and predict remaining useful life.
We have successfully applied this to filters, tires and mechanical components. Use small, inexpensive mems sensors and low-end microcontrollers to turn components into “smart” components.
> For more on sensing the environment around the vehicle
Edge AI solutions for Consumer and Commercial Products
Presence and Intrusion Detection
Use a variety of sensors to detect when someone is present nearby.
For use in security and facilities management applications. Built for the i.MXRT Anomaly Detection board from NXP, but can be ported to other platforms or sensor stacks.
Detect and classify specific motions. Sensitive enough for remote detection of breathing, chewing, falls.
Also useful for presence and intrusion detection. This Edge AI solution is coming soon, based on the Fujitsu Components FWM7RAZ01 doppler sensor.
Sound Event Detection
Identify common sound events for security applications such as glass breaking, babies crying, dogs barking, sirens and alarms etc.
Built for the i.MXRT Voice Assist and Anomaly Detection boards from NXP, but can be ported to other platforms.
Activity and Gesture Detection
Use accelerometers to identify specific motions, activities or gestures.
Built for “smart” personal care products but easily adapted to fitness and other applications, including User Interfaces. Works with Cortex M4, but depending on specific use case can be optimized for M3 or M0+.
Edge AI solutions for Industrial Use
Condition Monitoring and Predictive Maintenance
Identify the state of any machine in real time.
Requires customization to the specific machine type and function, but has been successfully used for pumps, conveyors, valves and motors - so far. Uses accelerometers, vibration sensors, and or sound, depending on the specific use case.
Remaining Useful Life and Time-to-Failure Prediction
Identify the state of wear and predict the time left before maintenance is required for mechanical components.
Requires customization for the specific use case. This Industrial Edge AI solution as been successfully used with valves, pumps, and filters - so far. Uses accelerometers, sound and/or differential pressure sensor, depending on the specific use case.
Detect unusual patterns on mechanical equipment, in real-time, using vibration, sound or other sensor inputs. Doesn’t just say that something is “vibrating more” -- it detects when it’s vibrating differently.
Requires configuration for specific use cases. Successfully used with pumps, fans, valves and other types of rotating equipment.
Your need not listed here?
Reality AI offers tools and services to help you create the exact Edge AI solutions for your product.
Get in touch with our team for more info.