Two builds of the same walking navigation app, measured side by side on a phone.
Outdoor augmented reality navigation is a tough test for a phone. The camera runs constantly, sensors feed the orientation, and anchors have to stay stuck to real places while the user walks. Developers building these apps face an early decision: use ARCore directly in native Android, or work through a game engine such as Unity.
A group of researchers put both approaches on the same footing. They built a location based AR prototype for outdoor pedestrian navigation twice, once in Unity and once as a native Android app, both powered by ARCore. Then they ran the two builds through the same outdoor scenarios and watched how hard each one worked the device.
The comparison focused on running load, the cost of keeping a live AR view alive while the phone also handles tracking and rendering. That load is what a user feels as heat, battery drain, and slowdowns when the camera pans across a busy street.
The native Android build came out ahead. It delivered the same augmented walking experience while asking less of the hardware, which is exactly the property that decides whether an app survives a long walk in the sun. The Unity build still worked, and it brought its own advantages.
The authors frame the result as a trade rather than a winner. The engine layer that makes Unity pleasant for 3D work also sits between the app and the device, and that shows up in running load once ARCore is tracking continuously outdoors.
The practical reading for teams is to pick by priority. If the navigation app is the product and it must run long sessions on modest phones, native Android is the safer base. If the project needs to ship widely and leans on heavy 3D assets, Unity earns its overhead.
Komang Candra Brata et al. (2023). An investigation of running load comparisons of ARCore on native Android and Unity for outdoor navigation system using smartphone. 2023 Sixth International Conference on Vocational Education and Electrical Engineering (ICVEE).
