A multimodal sensor attachment for a standard white cane that helps visually impaired users detect and avoid raised obstacles through haptic feedback.
While the basic white cane has served as a universally effective mobility aid for vision impaired users, it does have its limitations. By extending a cane during travel, users are able to detect objects approaching at the ground level. However, raised obstacles including hanging signs, tree branches and elevated railings remain dangerous. The Hummingbird Mobility Aid System attempts to fill this void by detecting raised obstacles and alerting users through haptic feedback.
The Hummingbird is a small, three mode sensor attachment that easily clips to any white cane. Press the mode button once and the Hummingbird activates its raised distance sensor. As a user approaches a raised obstacle, haptic feedback increases in frequency. A constant vibration in the handle indicates that the user should immediately correct their course. Press the mode button a second time and the Hummingbird enters line mode. Vibrations will persist until the user is within two meters of the person in front of them. When the line moves forward a vibrating motor will reactivate, indicating that the user should step forward until the vibration stops.
The Hummingbird is an ongoing personal project, conceived and developed by yours truly. Nonetheless, I would like to thank Andrew Davidson for his invaluable support and my brother, Christopher, for graciously letting me hijack his old canes.
The Hummingbird was developed in several stages. First, the system was breadboarded without a cane. This model was then used to test and refine the timing required for haptic feedback at the average walking speed.
To test the system on a white cane, a 3D printed case was created to house and attach the distance sensors. Hookup wire was used to separate the oversized Arduino and battery pack from the cane.
After experimenting with the original, three mode prototype, it was determined that the distance sensors were not accurate enough to sense slight elevation changes in mode three. To eliminate safety concerns, the mode was removed from the second version of the design. This also meant that the circuit could be streamlined by only using one distance sensor.
The second version of the Hummingbird was soldered to a small breadboard the size of an Altoids tin, cutting the Hummingbird’s form factor in half and preventing wires from coming loose during user testing.
Now that my Hummingbird prototype is sturdy enough for field testing, I intend to conduct a series of user tests with blind individuals outside of my immediate social circle (though my own brother has been immensely helpful as a pilot tester). Some of the questions that I still have left to answer include: 1) What is the ideal intensity for the hummingbird’s haptic feedback? 2) Does haptic feedback engage accurately and reliably for all users, especially those that navigate at great speed (high mobility)? 3) Are the Hummingbirds controls intuitive at on-boarding? If not, how might they be improved?
As design details are polished, I will look into creating a custom circuit board to minimize wiring and 3D print a more elegant cane clip/ sensor housing. Eventually a weatherproof housing will be required to test the Hummingbird in less than ideal environmental conditions (rain, cold, heat, wind, etc.).