We hypothesize that data-driven Functional Hemodynamic Monitoring and Machine Learning approaches on trauma patients in prolonged field care settings will enable domain experts to accurately characterize polytrauma events and appropriate therapies leading to accurate autonomous prolonged treatment interventions. Introduction: We describe results from a US Army funded research program called TRAuma Care In a Rucksack (TRACIR - Contract Award Number: W81XWH-19-C0101) that developed a closed-loop autonomous care resuscitation algorithm for non-compressible torso hemorrhage and evaluated it during combined air and ground evacuation in a traumatized porcine animal model. We show that a closed loop resuscitation algorithm – Resuscitation based on Functional Hemodynamic Monitoring (ReFit) - can effectively stabilize uncontrolled hemorrhagic shocked pigs in the laboratory and during both air and ground transport for 3 hours. The ReFit algorithm uses arterial pulse pressure and stroke volume variations to drive fluid and norepinephrine (NE) infusions to restore mean arterial pressure (MAP) and heart rate (HR) to clinically defined targets. Materials and Methods: An IACUC (Pitt) and ACURO (USAMRDC) approved study. Fifteen female swine were anesthetized, intubated and instrumented with femoral arterial and internal jugular catheters to measure MAP and infuse fluids and NE, respectively. A pulmonary artery catheter measured mixed venous O2 saturation (SvO2). A laparotomy exposed the liver. Following a 30-minute baseline the right lobe of the liver was lacerated twice with a through-and-through four-by-four cm stellate blade. Following the laceration, the animal was observed until MAP <40 mmHg for >5 minutes, then the liver was superficially packed, and abdomen closed. 30 minutes after reaching MAP <40mmHg ReFit started and ran autonomously for 3 hours. The first 8 swine validated ReFit algorithm in this model to reach MAP >60 mmHg and HR <110/min. Three control untreated swine died at ~60 minutes. Resuscitation started with 250 ml whole blood plus 1 g CaCl2, followed by a second 250 ml blood if still volume responsive, then lactated Ringers and NE as needed in 15-minute steps. The next 2 swine following blood infusion were assigned to air (medical helicopter flying a circuit) and the final 2 to ground by ambulance to a remote airfield and then air to hospital (n=2). After transport to hospital the swine returned to the laboratory for 3-hour total resuscitation time. Results: All twelve swine were autonomously resuscitated using ReFit algorithm to target MAP and HR within 45 min with 100% survival at 3 hours. The air only transport duration was 23- and 35-minute flight time and for ground to air 30 min ground plus 65- and 90-min flight time, respectively. ReFit restored SvO2 to baseline levels and showed a reduction of serum lactate, achieving enough documentation (TRL5) to proceed to pre-clinical studies. Conclusions: Our physiologically based autonomous ReFit algorithm effectively diagnosed and treated cardiovascular insufficiency due to severe uncontrolled hemorrhage independent of the site of care and during air and ground transport. The implications of these findings to care in austere or contested environments and during evacuation to a higher level of care are significant.