Onboard assessment of XRF spectra using genetic algorithms for decision making on an autonomous underwater vehicle

In order to optimise use of the limited resources (time, power) of an autonomous underwater vehicle (AUV) with a miniaturised X-ray fluorescence (XRF) spectrometer on board to carry out in situ autonomous chemical mapping of the surface of sediments with desired resolution, a genetic algorithm for rapid curve fitting is reported in this paper. This method quickly converges and provides an accurate in situ assessment of metals present, which helps the control system of the AUV to decide on future sampling locations. More thorough analysis of the available data could be performed once the AUV has returned to the base (laboratory).