Robust System Fault Classification & Detection System (RobustDetect) (RobustDetect)

NASA Glenn Research Center (GRC) is actively seeking improvements to fault detection and diagnosis, prognosis, fault recovery, and mitigation strategies to achieve greater system autonomy and resilience. For this, American GNC Corporation (AGNC) is proposing the Robust System Fault Classification & Detection System (RobustDetect) based on the observation that a classifier can only be as good as the quality of input data provided to it. The traditional neural network classification approach for fault diagnostics entails training the model on a variety of healthy and faulty operational states of a system to be monitored and deploying the model for real-time inference of system states given input sensor data streams. This assumes that the sensor data itself is an accurate measurement (of similar quality to the sensor data in the training set). However, transducers themselves can experience a myriad of faults (noise, drift, bias, etc.) that can be either intermittent or sustained. Using inherently incorrect data to make a conclusion about the state of a system being monitored is flawed as it will yield false positives/negatives regardless of classifier quality. The RobustDetect system addresses this problem by providing a sensor data validation algorithm that can be applied on sensor data to assign greater weight to sensors that are more likely to be healthy before being fed to a downstream classifier. The scheme involves a learning stage where a set of neural networks are trained to learn different relationships and correlations among the available sensors. Changes to these learned correlations can then be leveraged to determine if a given sensor is experiencing a fault itself and to either generate a virtual reading for that sensor (as a temporary self-healing strategy) or to assign less weight to that sensor within the classification pipeline. The result is a new approach to fault classification that is robust to sensors with degraded or faulty measurements.

The RobustDetect is focused on improving fault management capability for NASA systems by increasing the awareness of faults or off-nominal system behavior based on a diagnostic method that is robust to faults in the sensors themselves used to provide data to the FM systems. The target application would be NASA’s spacecraft such as CubeSats (e.g. 6U/12U/24U, where an example mission is the Lunar Flashlight that is requiring autonomous resilience to onboard errors) or other small satellites, the Europa Lander (to detect radiation-induced execution errors), Mars or Lunar Rovers (e.g. Mars Sample Return), future rotorcraft, the Orion Multi-Purpose Crew Vehicle, and Space Launch System. Because the system is not model based, it can be readily adapted to a variety of different systems and platforms (assuming the availability of training data for the neural network models). Medium-to-large missions with a low tolerance for failure can benefit from the robust sensor-validation technology, but also smaller missions are still applicable as the underlying technology is being developed to not be overly complex (given processing constraints in such missions). Other examples for NASA use include the Lunar Gateway or International Space Station. The system could be used for monitoring space habitat systems such as Environmental Control and Life Support Systems, Extravehicular Activity Systems, and Biological Life Support systems. The RobustDetect system could also be employed for assisting Integrated Systems Health Management (ISHM) of on-ground systems such as rocket engine testing facilities at Stennis Space Center, NASA Glenn Research Center’s Space Environments Complex (SEC) and thermal vacuum facilities, launch support systems at Kennedy Space Center (KSC), and more. The RobustDetect is a software-based fault management technology that can be readily adapted to a variety of companies and sectors outside of NASA. For example, it can be used by contractors involved in space systems development such as United Technologies Corporation, Honeywell International Inc., Lockheed Martin Corp, Northrop Grumman Corp, and Boeing. Additionally, private spaceflight companies including SpaceX, Virgin Galactic, Firefly Aerospace, Sierra Nevada Corporation, Bigelow Aerospace, etc. would benefit from the technology. However, the system is flexible and thus has the potential to be used in a variety of non-space applications including those of vehicles (e.g. aircraft, ships, robot platforms, rotorcraft, etc. as part of Integrated Vehicle Health Management Systems) and ground systems (e.g. such as machinery and equipment in chemical or materials processing plants, power generation and conversion facilities, oil and gas refineries, etc.). The RobustDetect is relevant to multiple sectors that are experiencing important growth, thereby demonstrating that the product has high commercial viability. Some examples include: (1) Small Satellites Market, where drivers include a growing demand for LEO-based small satellites (e.g. Starlink) and an increasing need for Earth observation imagery; (2) Fault Detection and Classification Market, which has recently experienced significant advances due to new AI/ML technologies, with software solutions having a major share of the market; (3) Machine Condition Monitoring, where a key driver for growth is predictive maintenance and industrial IoT; and (4) Supervisory Control and Data Acquisition (SCADA) where increasing advances in wireless sensor networks is expected to contribute to the expansion of this market.

This is a mature technology (TRL 7+) — the realistic path in is usually NASA's Technology Transfer Program: licensing an existing NASA patent, or a Space Act Agreement to use NASA facilities/expertise directly. NASA also runs a startup licensing program with no upfront fee for companies formed to commercialize a specific NASA technology.

None of these are guaranteed paths for this specific project — TechPort itself doesn't have an "apply" button. Reaching out to the contact(s) above with a specific question is usually the fastest way to find out what's actually open.