There is interest for military personnel to be equipped with wearable, wireless blast gauges that measure overpressure magnitudes from blast explosions with respect to time. These blast gauges are typically worn on the chest, shoulder, and back of helmet. Controlled experiments have shown that the blast gauges can accurately measure the blast overpressure as compared to pencil gauges. However, at best, these sensors measure pressure at the location they are worn, when what is needed is the overpressure that propagates inside the head. However, there has been limited exploration of how to develop the “sensor-to-head” transfer function. In this study, various methods to establish this transfer function are examined. One approach is by using triangulation of the of the chest, shoulder and helmet pressure traces to identify the magnitude and location of the blast source – the so-called inverse-source localization. The nature of triangulating the data is complex due to the variety in explosives as well as the non-ideal and compressed gases present. This thesis explores triangulation using two methods, geometric algorithm and machine learning, to analyze data from CTH and Viper computational models.
