The student about to move to the city
A blind teenager is about to leave home for his undergraduate degree, and the city ahead of him is daunting. New routes, new hazards, none of them memorised yet. Every unfamiliar space charges the same cost: the mental load of holding a map he cannot see, the time each basic task takes, the quiet drop in confidence when the built world was not designed with him in mind. His white cane reads the ground a step ahead and nothing of the world past it. Lead started from his morning, not from a gadget.
Mapping the whole problem
I did not start by choosing a solution. I started by mapping the entire problem space of visual impairment, wide and on purpose, on a shared board that grew to dozens of life domains: finance, healthcare, cooking, emergencies, ageing, ATMs, navigation. Going that broad was the point. It let me see the correlations between segments before committing to any one of them, and find the gaps nobody had addressed.
That map sorted into three research areas. Braille and education, where the shortage of scribes pushes students out (our research cited 77 of 140 students discontinuing their studies for lack of a scribe, and a single page becoming three in Braille). Daily living skills, the cognitive load of tasks like identifying clothes by texture and stitching. And orientation and mobility, moving through space, where the stakes are highest because, as the research put it, eighty percent of our spatial understanding is visual.
Narrowing to mobility
I converged on orientation and mobility and set the other two aside, documented but not chosen. The insights that pointed there were specific. Blind people form chains, one person with some vision leading the rest, which told me the answer was social as much as technical. Users need about a metre of buffer to walk without fear. They stop using a cane in familiar places once the space is memorised. And they are reluctant to wear glasses however well designed, which quietly ruled out every head-worn idea.
The brief that came out of it: give visually impaired teenagers better ways to get past obstacles on their own, and connect them to a community that is motivated to help.
Designing the foldable cane
This is where the project became an object, not just an idea. Before drawing anything I studied cane typologies, the short, floor-gripping support cane against the long probing cane, and the family of tips that already exist: slidable, roller, pencil, marshmallow, metal guide. The cane I designed is telescopic, with distinct working modes: collapsed for crowds, extended for the open street, and a sensing mode for traffic. Telescopic on purpose, so it works in tight spaces and folds away when it is not needed.
The body is organised as a stack: a rechargeable battery station at the base, a circuitry station mounted above it, a pan-and-tilt camera in the shaft that orients itself as the user moves, and the smart tip at the working end. The tip took a tortoise as its cue, the sensor tucking in and out of a protective housing. Materials were chosen against real wear, a light polyethylene shaft and a tip explored in ceramic, nylon, and metal. Audio came back through bone conduction, so the ears stay open to the street.
The system underneath
The electronics were built on Arduino: an ultrasonic ranger for distance, a gyroscope to track orientation, a light sensor for unsafe low-light areas, a buzzer, and a Bluetooth link to the phone. The more ambitious half ran on a Raspberry Pi with a camera, using open-source object detection to name what was ahead, with an on-device AI stick to keep the inference local. The phone did the talking, turning the cane’s stream into calm audio directions.
The idea I am still fond of was pattern recognition. Different terrain leaves different signatures in the sensor stream, and I mapped distinct patterns for a staircase, a ramp sloping up or down, a level transition, and a ditch. It pointed at a cane that learns the shape of the world rather than only the distance to the next thing.
The community half
A device is not enough on its own, and the chain insight said so. The second half of the system was a volunteer service: an app where sighted volunteers register to help, live video assistance when a user is stuck, and a gentle bridge to eye and cornea donation. A cane that senses, and a network that sees.
Building and testing it
The prototype was honest about being a prototype: a breadboard thick with wires, a repurposed telescopic monopod standing in for the cane shaft, sensor readings streaming live to an Android phone. I ran empathy tasks at the institute, reaching a reading section, finding an object on a desk, getting a glass of water, and walked real routes around Yeshwantpur and the Peenya metro, logging every hazard the cane would have to catch: dripping AC vents, poles on the footpath, faulty tactile tiles, a slope with no warning.
What I dropped
The honest part is what did not survive. DIY bone-conduction earphones I built and abandoned for poor sound. The pan-and-tilt camera and the object detection stayed work in progress, kept as the version-two ambition rather than dressed up as finished. Naming the dead ends is the point: they are how a sprawling problem map narrowed to a cane, a tip, and a community that could actually be built.
What it taught me
Lead was never awarded or deployed, and I am not going to dress it up as more than it was. What it gave me is the method I still use: map the whole problem before you fall in love with a solution, converge on the one worth solving, then design the real object all the way down to its materials and its failure modes. Years before I did this for a living, this is where I learned that the research is the design.