AI-Powered Multimodal Assistive Device for Enhanced Accessibility
19/813,182
- Nonprovisional utility patent application
- 19/813,182
- 3286
- U.S. Patent and Trademark Office
- 20 September 2026
- Not yet publicly available
Filed under 35 U.S.C. 111(a). Pending; not yet published or examined.
Sensors
Capture information about the user's surroundings.
Generative AI
Interprets the surroundings and emotional cues and decides what is relevant to communicate.
Haptic feedback
Translates that information into structured vibration patterns the user can feel.
- Smart cane
- Smart glasses
- Wristband
- Glove
The problem it addresses
Much of the information people use to move through the world, from obstacles and layout to the expressions of the people around them, is visual. For blind and visually impaired people, that information has to arrive another way.
This invention explores communicating it through touch: environmental information and emotional cues are translated into structured vibration patterns, so that a user can feel what is around them instead of hearing a stream of spoken descriptions.
How the system is intended to work
The design brings three parts together. Sensors capture the user's surroundings. Generative AI interprets what the sensors capture, including emotional cues, and determines what is relevant. Haptic feedback then expresses that interpretation as structured vibration patterns.
The same approach can take different physical forms: the application describes potential implementations as a smart cane, smart glasses, a wristband or a glove.
Support and next step
Varish was awarded a $1,800 micro-grant by the Managing Partner of Wyser VC to develop an AI-powered haptic smart cane, the first form factor being explored.
Where it stands
This is an invention described in a pending U.S. patent application, filed on 20 September 2026. It describes intended functionality: the device has not been clinically validated or deployed, and no outcome claims are made here.