Online Track Reconstruction Method Based On An Artificial Neural Network For Gas Detectors

Micro pattern gaseous detectors(MPGD)are widely used for radiation measurements in many fields including particle physics experiments and atomic nuclear physics experiments,especially for applications requiring large-area high-precision position measurements,due to their short signal collection time,high position resolution accuracy,and good radiation resistance.In such applications,the detector requires a large number of readout electrodes,so its electronics and data acquisition system face difficult technical challenges.In addition,the measurement of the incident particle position by the conventional center of gravity method inevitably introduces systematic errors,which is an urgent problem to be solved for further improvement of the detector position resolution.In this paper,an artificial neural network algorithm-based online track reconstruction scheme for micro pattern gaseous detectors is proposed.Firstly,Geant4 software is used to simulate the generation of detector signals for the training and testing sample sets of two cascaded neural networks.After the training of the algorithm,the data are analyzed using the traditional center of gravity method and the trained and shaped network,respectively.The results show that the method has higher accuracy and efficiency compared with the center of gravity method.In addition,this paper presents a scheme to implement the algorithm on a field-programmable gate array chip to demonstrate that the algorithm can be integrated into modern data acquisition systems used in online imaging techniques.