Sensors drift, light changes, and the interface still has to point the right way.
An augmented reality walking guide sounds like a rendering problem. Most of the difficulty sits underneath, in knowing where the user is, where the target is, and how far to trust the sensors in between.
BusAR was built in three layers. The phone handles the camera and the overlay, a middle layer holds the guidance logic, and web services supply stop locations and route data. Keeping them apart meant sensor handling could change without touching the interface.
Location comes from GPS. Heading needs the magnetometer, which drifts near metal and among tall buildings, so the accelerometer and gyroscope help steady it. Distance to each stop comes from the great circle formula.
The app was tested on real streets rather than in a controlled setup, which is the only way to learn whether a walking guide holds up. Sensor drift and shifting light are exactly what a desk test cannot reproduce.
The efficiency result is encouraging for adoption. Around 0.25 kilobytes per second is light enough for a limited data plan, and it came in lower than the route planning apps used for comparison.
The design lesson is to spend effort where the user actually gets confused. A list of stop names is accurate and useless when you are on the wrong side of the road, while an arrow drawn over the camera answers the only question that matters.
Brata, Liang, and Pramono (2015). Location-Based Augmented Reality Information for Bus Route Planning System. International Journal of Electrical and Computer Engineering.
