Orbit Logic has been awarded a Phase I Small Business Innovation Research (SBIR) contract sponsored by the National Aeronautics and Space Administration (NASA) to develop the Fault Learning Agent for Prediction, Protection and Early Response (FLAPPER) System – a satellite solution to utilize Machine Learning (ML) to autonomously detect, isolate and correct on-board faults. FLAPPER’s goal is to significantly reduce (and in time, eliminate) the human resources expended in the traditional process of detection and correction of spacecraft faults. Additionally, this would improve responses for missions with limited communications. Reducing the time spent by operators to analyze the root cause, and even providing the capability for the satellite to quickly return to operational state are both seen as obvious benefits of the technology.
The FLAPPER solution will expand our onboard Autonomous Planning System (APS) architecture to include ML capable of detecting, isolating, and mitigating anomalies in real- or near-real-time with minimal ground intervention. A set of defined fault detection and correction constraints will be developed, along with the capability for operators to classify new types of faults. These constraints, along with spacecraft data input, will be used to train a Specialized Autonomous Planning Agent (SAPA) called the FLAPPER Detect SAPA - to detect and classify faults based on novel telemetry limits and value trends. The corrective component of the FLAPPER system (FLAPPER Correct SAPA) will then plan correlated corrective actions to mitigate each fault type.