The emerging field of radio based neutrino astronomy holds promise in answering long lasting questions about our universe such as identifying the sources of ultra-high energy (UHE) cosmic rays. This requires a multi-messenger effort in astronomy which is a relatively new collaboration between the various particles/waves used to study space. It is important for radio neutrino astronomers to show that UHE neutrino detection can be made with excellent precision in direction and energy reconstruction for the field to grow from pilot phases into large scale experiments. The Antarctic Ross Ice-shelf Antenna Neutrino Array (ARIANNA) is one such detector with this goal. ARIANNA aims to detect UHE neutrinos via radio (Askaryan) emission from particle showers when a neutrino interacts with ice, which is an efficient method for neutrinos with energies between 10^16 eV and 10^20 eV. The ARIANNA radio detectors are located in Antarctic ice just beneath the surface. Neutrino observation requires that radio pulses propagate to the antennas at the surface with minimum distortion by the ice and firn medium. To reconstruct the direction, a measurement of polarization, radio frequency signal direction, and viewing angle off of the Cerenkov cone, along with ice attenuation and signal trajectories must be made. The energy further requires a modeling of the Askaryan radiation created from the stochastic processes of particle showers in dense media. This results in an irreducible energy resolution, which sets the goal for energy reconstruction techniques.

An experimental evaluation of radio signal polarization and direction resolution was completed using the residual hole from the South Pole Ice Core (SPICEcore) Project. Radio pulses were emitted from a transmitter located down to 1.7km below the snow surface. After deconvolving the raw signals for the detector response and attenuation from propagation through the ice, the signal pulses show no significant distortion and agree with a reference measurement of the emitter made in an anechoic chamber. The origin of the transmitted radio pulse was measured with an angular resolution of 0.37 degrees, indicating that the neutrino direction can be determined with good precision if the polarization and viewing angle of the radio-pulse can be well determined. In the present study we obtained a resolution of the polarization vector of 2.7 degrees. Neither measurement show a significant offset relative to expectation. We also report on the results of a simulation study of the ARIANNA neutrino direction and energy resolution. The software tool NuRadioMC, which is rapidly becoming the industry standard, was used to reconstruct the polarization and viewing angle to determine the neutrino direction. Multiple models of Askaryan radiation and detector sites along with a range of neutrino energies were tested. The neutrino space angle resolution was determined to be below 3 degrees, which is governed by the polarization uncertainty of the same scale. The polarization reconstruction from experimental SPICEcore studies showed large systematic errors that results in the 2.7 degrees uncertainty, whereas the per depth resolution is a sub degree statistical error. Therefore it is expected that the polarization resolution, which is the dominant contribution to the neutrino space angle resolution, will be greatly improved in future studies by determining and eliminating systematic effects such as antenna modeling. Neutrino energy resolution is reported at 40% for 10^18 eV neutrinos, which is below the inelasticity limit and therefore ARIANNA is not limited by its detectors ability to reconstruct the energy.

The ARIANNA experiment is a radio Askaryan detector located at the Ross Ice Shelf in Antarctica, designed to measure radio signals generated by high-energy neutrinos ($E_{\nu} > 10^{17}$ eV) interacting with Antarctic ice. The primary goal of this work is to develop new hardware and software that enhance the overall sensitivity of ARIANNA stations, making them more sensitive to these neutrino interactions. This thesis comprises four individual projects: reducing noise from the Battery Management Unit (BMU), developing a trigger bandwidth filter, implementing deep learning (DL) algorithms for real-time event filtering, and designing and testing a next-generation low noise amplifier (LNA).

The first project focused on identifying and mitigating the BMU noise that interfered with ARIANNA data. As the primary power manager for each ARIANNA station, the BMU would "flip" between states during periods of low sunlight, causing significant noise pulses to propagate through the power lines and trigger the station at regular intervals. A solution was proposed to reduce the frequency of these triggering events, minimizing unnecessary memory usage and processing time. The remaining three projects aimed to enhance the overall sensitivity of the stations. The development of a trigger bandwidth filter would enable the stations to respond to signals in the range of 100 MHz to 300 MHz (the peak of Askaryan radio power) while still digitizing a wider range of frequencies between 80 MHz and 500 MHz of the same input signal. This improvement could increase sensitivity to neutrinos with energies between $10^{17}$ eV and $10^{18}$ eV by up to 1.5 times the base sensitivity.

Additionally, the DL algorithm was implemented to assess whether ARIANNA stations could, in real-time, filter out non-neutrino-like events Reducing the data that needs to be analyzed for actual neutrino events and prioritizing memory allocation for neutrino-like events. This would not only streamline data processing but also allow for lower station thresholds. Consequently, this reduction in thresholds would enhance the station’s sensitivity to neutrinos with energies above $10^{18}$ eV. When combined with the newly designed LNA, the overall sensitivity of the stations could improve by at least 1.5 times for energies ranging from $10^{17}$ eV to $10^{20}$ eV, excluding the advancements from the bandwidth trigger. The process of developing a new LNA is also presented in this work.