Despite receiving oxygen therapy, many COPD patients experience extended periods of hypoxemia during routine daily activities. In others, inappropriately high oxygen flow rates can depress hypoxic drive leading to worsening CO2 retention. As flow-rates during LTOT are fixed, oxygen delivery will not respond to patients' fluctuations in oxygen demand. The research project has aimed to develop and evaluate a closed-loop control method capable of actively varying flow-rates in response to the measured oxygen demand. We demonstrate how SpO2 from ambulatory or overnight pulse oximetry can be used as feedback into an automated O2 flow-rate controller. A model to mimic the patient oxygen saturation response has been developed in a computer simulation to help characterize the closed-loop system. With the collaboration of the Academic Unit for Sleep and Breathing at the Royal Brompton Hospital, the controller response has also been validated against patient saturation measurements recorded during overnight pulse oximetry monitoring. Preclinical computer simulations indicated an improved matching between oxygen supply and demand, maintaining SpO2 above threshold to maximize therapeutic efficacy. An investigational system capable of regulating the Saturation Driven Oxygen Therapy (SDOT) was constructed. In a randomised cross-over clinical pilot study, we further evaluated the SDOT system against constant-flow LTOT during exercise. The clinical results indicate that compared to standard oxygen therapy, SDOT produced a significant reduction in time spent with hypoxemia, decreased the extent of hypoxemia and lowered mean heart rates during exercise. Moreover, for patients with acceptable resting oxygen levels, SDOT provided conservation benefits by reducing the rate of oxygen consumption. The study established the potential to significantly improve the efficacy and economic delivery of this gold standard therapy.