Modern rebreathers rely on galvanic oxygen sensors to monitor the composition of the breathing gas to enable humans to safely breathe for long periods with minimal supplemental oxygen. While these sensors are effective, they are prone to failure and typically exhibit poor noise characteristics. In this paper a Kalman filter based approach is evaluated for the purposes of improved rebreather state estimation. A Kalman filter based approach is proposed due to its ability to accurately estimate system states in the presence of substantial sensor noise. Testing and simulation procedures are presented and the author concludes that the use of a Kalman filter effectively filters out sensor noise. In addition, an augmented Kalman filter is presented with a discussion of sensor failure detection capabilities.