Modern, robotic 'white cane' brings navigation assistance to the 21st century

Equipped with a colour 3D camera, an inertial measurement sensor, and its ain on-board computer, a recently improved robotic cane could connection unsighted and visually impaired users a caller mode to navigate indoors. When paired with a building's architectural drawing, the instrumentality tin accurately usher a idiosyncratic to a desired determination with sensory and auditory cues, portion simultaneously helping the idiosyncratic debar obstacles similar boxes, furniture, and overhangs. Development of the instrumentality was co-funded by the National Institutes of Health's National Eye Institute (NEI) and the National Institute of Biomedical Imaging and Bioengineering (NIBIB). Details of the updated plan were published successful the diary IEEE/CAA Journal of Automatica Sinica.

"Many radical successful the visually impaired assemblage see the achromatic cane to beryllium their champion and astir functional navigational tool, contempt it being century-old technology," said Cang Ye, Ph.D., pb writer of the survey and prof of machine subject astatine the College of Engineering astatine the Virginia Commonwealth University, Richmond. "For sighted people, technologies similar GPS-based applications person revolutionized navigation. We're funny successful creating a instrumentality that closes galore of the gaps successful functionality for achromatic cane users."

While determination are compartment phone-based applications that tin supply navigation assistance—helping unsighted users enactment wrong crosswalks, for example—large spaces wrong buildings are a large challenge, particularly erstwhile those spaces are unfamiliar. Earlier versions of Ye's robotic cane began tackling this occupation by incorporating gathering floorplans; the idiosyncratic could archer the cane whether helium oregon she wished to go, and the cane—by a operation of auditory cues and a robotic rolling tip—could usher the idiosyncratic to their destination. But erstwhile utilized implicit agelong distances, the inaccuracies successful the user's determination could physique up, yet leaving the idiosyncratic astatine an incorrect location.

To assistance close this issue, Ye and colleagues person added a colour extent camera to the system. Using infrared light, overmuch similar a mobile phone's front-facing camera, the strategy tin find the region betwixt the cane and different carnal objects, including the floor, features similar doorways and walls, arsenic good arsenic furnishings and different obstacles. Using this information, on with information from an inertial sensor, the cane's onboard machine tin representation the user's precise determination to the existing architectural drafting oregon floorplan, portion besides alerting the idiosyncratic to obstacles successful their path.

"While immoderate cell phone apps tin springiness radical auditory navigation instructions, erstwhile going astir a country for example, however bash you cognize you've turned conscionable the close amount?" said Ye. "The rolling extremity connected our robotic cane tin usher you to crook astatine conscionable the close constituent and precisely the close fig of degrees, whether it's 15 degrees oregon 90. This mentation tin besides alert you to overhanging obstacles, which a modular achromatic cane cannot."

There are inactive a fewer kinks to beryllium worked retired earlier the strategy volition beryllium market-ready—it's inactive excessively dense for regular use, for example, and Ye's squad is looking for a mode to slim down the device. Nevertheless, with the quality to easy power betwixt its automated mode and a simpler, non-robotic "white cane mode," Ye believes this instrumentality could supply a cardinal independency instrumentality for the unsighted and visually impaired, without losing the characteristics of the achromatic cane that person stood the trial of time.

---

---

More information: He Zhang et al, An RGB-D Camera Based Visual Positioning System for Assistive Navigation by a Robotic Navigation Aid, IEEE/CAA Journal of Automatica Sinica (2021). DOI: 10.1109/JAS.2021.1004084