Chemical Detection Using the Receptor Density Algorithm

Abstract

This paper describes the application of the receptor density algorithm, an artificial immune system, as used to detect chemicals from data provided by various spectrometers. The system creates chemical signatures which are matched to a library of known chemicals, allowing the positive identification of hazardous substances. The performance of the system is tested against a publicly available mass-spectrometry dataset, against which it has previously been demonstrated as an effective anomaly detection algorithm. An autonomous chemical-detection device is then discussed, in which the algorithm is running on hardware embedded in a Pioneer robot carrying a portable chemical agent monitor.