Edge device project
Mobility Assistant
A custom edge device that detects objects and presence in dynamic environments, running compact machine-learning models directly on single-board hardware.
How it senses and responds
Camera and sensor inputs
Cameras and sensors capture the surroundings the device must read, matched to the conditions where it operates.
Object recognition models
Compact models identify objects and their presence, trained for the environment the device will actually face.
Feedback systems
Haptic, visual, or audio alerts give the user a timely signal to support real-time decisions.
On-device processing
Low latency without the cloud
Models run directly on the single-board hardware, so detection responds without waiting on a network round trip. That keeps the device useful where connectivity is unreliable and autonomy matters most.
