This project, Heartbeat (Heliosphere to Earth Atmosphere Rendering Through Building Excellent Artificial-Intelligence Training) involves studying and forecasting phenomena in the Earth's upper atmosphere and ionosphere at high latitudes associated with space weather. In response to solar and geomagnetic storms, the ionosphere undergoes drastic changes and becomes highly irregular. This can lead to disruptions in radio communications and related technologies such as radio navigation (GPS) and imaging. This DARPA award to Georgia Institute of Technology and its partner institutions— including Cornell—is fostering nontraditional means of forecasting the disruptions utilizing novel data sets and machine-learning algorithms.

David Hysell, Earth and Atmospheric Sciences, and his group are responsible for supplying ionospheric measurements. Hysell is deploying a network of high-frequency radio transmitters and receivers in Alaska. The signals passing between the transmitters and receivers employ pseudorandom noise codes such as those used with GPS signals. The received signals provide parameters that characterize the electron number density in the ionosphere through which they pass.

Using statistical inverse methods, Hysell is reconstructing, from the signals, a regional specification of the ionosphere nearby. Disruptions and irregularities in the ionosphere will be clear in the data. The intent is that these measurements can be used to train machine-learning algorithms to anticipate disruptions and to provide accurate forecasts of space weather.