Predicting Astronaut Radiation Doses From Large Solar Particle Events Using Artificial Intelligence

For deep space missions, a major concern is the occurrence of large solar particle events. In this work a dynamic, new type of artificial neural network called a Sliding Time Delay Neural Network that is capable of accurately predicting total dose for an event, from several input doses early in the event, is presented. The network can update its total dose predictions during the event as new input data are received. Results from testing indicate that the network can predict total doses from large events that are outside the training set to within 4% very early in the event.