A Geometry-Independent Framework for Charged-Particle Track Reconstruction in the Mu2e Experiment

The Mu2e experiment at Fermilab searches for charged lepton flavor violation via the neutrinoless conversion of a muon into an electron in the field of an aluminum nucleus. Achieving the required single-event sensitivity demands precise reconstruction of charged-particle tracks and energy deposits within the straw-tube tracker. We present a preprocessing and foundational machine learning–based reconstruction framework that infers hit timing, position, and energy directly from digitized tracker readouts. By learning event-level correlations among straw signals, the model achieves robust millimeter-scale position reconstruction and sub-nanosecond timing precision, without explicit geometric inputs. This geometry-independent, data-driven approach establishes a foundation for improved track and energy reconstruction in Mu2e and offers a scalable paradigm for next-generation high-rate tracking detectors in charged-lepton-flavor-violation experiments.