Iterative Track Fitting Using Cluster Classification in Multi Wire Proportional Chamber

This paper addresses the problem of track fitting of a charged particle in a Multi Wire Proportional Chamber (MWPC) using cathode readout strips. When a charged particle crosses a MWPC, a positive charge is induced on a cluster of adjacent strips. In the presence of high radiation background, the cluster charge measurements may be contaminated due to background particles, leading to less accurate hit position estimation. The Least Squares method for track fitting assumes the same position error distribution for all hits and thus loses its optimal properties on contaminated data. For this reason, a new robust algorithm is proposed. The algorithm first uses the known spatial charge distribution caused by a single charged particle over the strips, and classifies the clusters into "clean" and "dirty" clusters. Then, using the classification results, it performs an iterative Weighted Least Squares fitting procedure, updating its optimal weights each iteration. The performance of the suggested algorithm is compared to other track fitting techniques using a simulation of tracks with radiation background. It is shown that the algorithm improves the track fitting performance significantly. A practical implementation of the algorithm is presented for muon track fitting in the Cathode Strip Chamber (CSC) of the ATLAS experiment.