The goal of the ARCANA project is to produce a road condition sensor that can be used in road weather information systems. ARCANA implements signal processing artificial neural network (ANN) technology to classify signals recorded with a microphone. This paper is a review of ANN technologies and their application to ARCANA. The entire process is discussed from data acquisition, feature extraction, and ANN design to the actually training of the network. Preliminary classification results in an independent validation set yielded 88% correct classification. This was improved to a full 100% with the addition of a confidence level threshold.