Fast centroid estimation algorithm for high-rate detectors based on a two-point Gaussian fit

Recent developments in detector back-end electronics for position sensitive gas proportional detectors have moved the data processing bottleneck out of the front-end electronics and into signal processing. Even with today's fastest DSP's, the centroid estimation of the cathode charge distribution can impose a limitation on the ultimate count rate attainable by the detector. Here we present a centroid-finding algorithm that does the bulk of the centroid calculations based on a two-point Gaussian fit that assumes a priori knowledge of the distribution's variance. The variance is estimated only in the calibration phase. The two-point algorithm has the advantage that no division is required, greatly reducing the computational burden on the DSP. The performance of the algorithm is evaluated by implementation on existing detectors.