New Techniques in Sensor Waveform Processing for the IceCube Neutrino Observatory

Future extentions of the IceCube Neutrino Observatory will include many strings of digital optical modules (DOMs), each with multiple photomultiplier tubes (PMTs). In order to reduce the bandwidth requirements of the deep-ice cable, part of the processing of the PMT voltage waveforms will occur in the DOMs instead of sending all raw voltage waveforms to the surface. We demonstrate that implementing the first step of the Lawson-Hanson non-negative least squares algorithm on a microcontroller unit in the DOMs can process dark noise hits by unfolding the waveform using the PMT single photoelectron response. This single photoelectron (SPE) processing algorithm can be accelerated substantially by incorporating additional SPE time information. In addition, we present a new voltage waveform processing algorithm utilizing neural networks that was developed to potentially accelerate voltage waveform processing.