Abstract: Systems and methods for detecting faults in estimator data are provided. In one or more examples, measurements from one or more data sources is received. The received data is used to calculate an innovation which in one or more examples can represent the difference between the received data and an expected value of the data. The innovation can then be used to generate a test statistic which is then accumulated at a plurality of monitors, wherein each monitor of the plurality of monitors is configured to accumulate the test statistic over varying periods of time. The accumulated test statistic at each monitor can be compared against a predefined threshold that is set for each individual monitor. If the accumulated test statistic at any particular monitor is found to be above its corresponding predefined threshold, then the system can alert the user that the received data is likely to be faulty.