Sensor data fusion approaches to understanding nitrate at higher resolution in coastal ocean areas

Chemical sensing of natural waters at a spatially relevant resolution relevant demands sensors that (1) are relatively inexpensive, (2) capture holistic characteristics of variable chemical background, and (3) are able to detect target pollutants that are a minority fraction of the charge balance.  Ion-selective electrodes are promising due to their small size, low power needs, wide variety of targets, and ability to be manufactured at low cost, but the issue of imperfect selectivity combined with variable chemical backgrounds points to the need for continued innovation in this space.  This research expands on previous sensor fusion (“electronic tongue”) efforts through a combined software/hardware design, including selection of different membrane compositions to offset known interferences.  This talk will present the design framework and preliminary results including customizable screen-printed ISE arrays, 128-channel data acquisition circuitry, and a sequential injection analysis (SIA) system for automatically generating machine learning algorithm training data for fresh to saline samples.