Both autopilot systems and human pilots, particularly human pilots operating in instrument meteorological conditions, rely heavily on sensor feedback to safely control aircraft. The loss of reliable information for even a single state feedback signal can easily initiate a chain of events that leads to an accident. Even when hardware redundancy is employed, common-mode failures are a significant hazard that can make hardware redundancy ineffective for achieving the desired system reliability. For example, multiple pitot tubes can experience a common-mode failure during an icing event, depriving the pilot of vital airspeed information. The proposed virtual redundancy approach can significantly improve flight safety by identifying failed sensors and estimating the correct output values as replacements for those failed sensors. Estimates are based on a rigorous statistical formulation that makes optimal use of all available information including feedback from all remaining physical sensors, nonlinear models of vehicle dynamics, and models of actuator and sensor responses. The proposed research will also develop strategies for enabling pilots to make effective use of the virtual sensor outputs, including guidance algorithms that identify a trajectory that maximizes the likelihood of maintaining safety of flight and cueing techniques that allow the pilot to follow the resulting trajectory while minimizing the increase in workload.