Automated Measurement and Collection of Crowd-Sourced Weather Observations Using Smartphones

Latest generation smartphones now include onboard sensors for direct measurement of environmental conditions such as ambient temperature, relative humidity, and barometric pressure.  Other available sensors such as the light sensor, camera, microphone, and inertial measurement unit could potentially be used to estimate additional surface conditions such as solar radiation, precipitation, wind speed, and wind direction.  Exploitation and assimilation of such real-time observations from broadly distributed and spatially dense networked mobile devices could lead to improved awareness of current weather conditions and more accurate forecasting at regional, national, and global scales.  However, measurement and collection of these observations should ideally require no user interaction and not interfere with the primary intended uses of the mobile device.

This paper evaluates the feasibility of using the full suite of onboard mobile device sensors for automated measurement and collection of crowd-sourced weather observations.  We first describe several techniques for direct and indirect estimation of weather conditions, including prior published approaches by others in this area.  We then assess different mitigation strategies such as crowd-sourcing, geospatial location analysis, machine learning, and physics-based modeling to address confounding factors such as device motion and orientation; internal device heating; sensor noise; sheltering effects due to clothing, vehicles, and buildings; and physiological heating and sweating.  We then describe a prototype cloud computing server architecture for collecting, processing, archiving, and publishing of smartphone-based crowd-sourced weather observations.  We next give a live demonstration of our prototype smartphone-based application interfaced to our prototype cloud server.  Finally, we discuss potential applications of this information, including data assimilation for improved regional severe weather forecasting, and crowd-sourced situational awareness of weather and terrain conditions in the battlespace using soldier-worn mobile devices.