New position sensing method for position-sensitive proportional counter with neural network algorithm

Abstract We have developed a new position sensing method that can obtain position information directly from the pulse shape of a position-sensitive proportional counter (PSPC). A digital signal processing technique and a neural network algorithm have been used to recognize the pulse shapes and to obtain position information. The method has been applied to a one-sided read-out type PSPC where the pulse shapes depend strongly on the position of radiation interactions. The neural network can recognize the pulse shapes and provide the proper position even for positions that have not been taught in the learning process. The best relative position resolution is 6.4 × 10 −3 . Fairly good integral linearity has been obtained throughout the effective length of the PSPC.