A German AI company is developing a machine learning system to help blind and visually impaired people find their way safely around towns and cities.
AIServe was founded in 2017 by CEO Gustavo Madico to combine computer vision, artificial intelligence (AI), and wearable hardware to make walking safely and independently accessible to everybody, via simple devices that the user talks to, and receives instructions from.
The new device would combine AI and location services with a camera, computer vision system, microphone, and earphones.
The market is certainly large and underserved, says the company. Globally, 253 million are thought to be living with vision problems, according to the World Health Organization, with blindness affecting 36 million.
AiServe estimates that one in three seniors will suffer from a visual impairment that prevents them from navigating safely, reinforcing recent research which says that aging people in aging cities will be the big challenge for Western economies over the next 20 years. According to the WHO, 81 percent of blind or visually impaired people are aged over 50.
In the US, an estimated 7.3 million people are affected by sight loss. If 14-18 percent of those are blind or partially sighted, then that equates to over one million citizens.
In the UK, over two million people are thought to be affected by sight loss, with 360000 (18 percent of visually disabled people) registered as blind or partially sighted. By 2050, the number will soar to nearly four million visually impaired citizens, according to the Royal National Institute for the Blind, as the population ages and problems such as diabetes and obesity rise.
According to AiServe’s research, estimates of nearly 70 percent of the $40 billion assistive technology market in the US alone is dedicated to visual impairment aids, few of which help people get from A to B safely.
The company has identified a sizable market opportunity to build a wearable machine learning system that, in the company’s words, ‘learns how to walk like a human‘ not by aping human mobility, but by learning to identify the visual cues that sighted people recognize easily in cities, such as buildings, paths, street furniture, pavements, sidewalks, curbs, and corners.
As the system learns over time by acquiring more and more data, the computer vision system will be able to map city blocks and neighborhoods and give satnav style-instructions to the wearer, such as “Walk straight ahead for 100 feet.”
AiServe plans to release production models in the UK and Germany by next summer.