Abstract: A novel system and method for detecting and recovering from single event upsets in integrating processing systems are presented. Two delta value processing lanes may be utilized to process raw data from inertial sensors. The output of the two processing lanes may be compared to detect the presence of an SEU. If an error is detected, a previously stored set of delta values may be utilized by the system. Additionally, the system may contain a temporary buffer to store calculated delta values in the event of a processor reset caused by an SEU or other error. The temporary buffer may store delta values accumulated during the processor reset and may furnish these stored delta values to the processor once it has finished resetting. The stored delta values may then be batch-processed and applied to system state values saved to memory prior to the processor reset.