Alpha–Gamma Discrimination in BaF2 Using FPGA-Based Feedforward Neural Network

Abstract

In this paper, we investigated the hardware implementation of two feedforward neural networks (NNs) using a field-programmable gate array (FPGA), then used the networks for alpha–gamma discrimination in barium fluoride (BaF2). The BaF2 detector output was sampled using a 1-GSPS ADC, and then we extracted six information of the pulses in FPGA as the input features to the NNs. The performance of this method turned out very good, the false alarm rate of the networks was less than 0.3%. Besides, dead time of the networks was less than 820 ns. Low logic occupancy is also discussed.