Sources of sensor error leave a UTM (UAS Traffic Management) system vulnerable. The impact of anomalous sensor behavior can be devastating to a UTM system. Degraded sensor accuracy can lead to broken tracks or split tracks within a UTM system. Degraded sensor sensitivity may result in a Mid-Air Collision with an aircraft not detected by UTM sensors. It may also result in more Loss of Well-Clear violations and more frequent Near Mid-Air Collisions.
This proposal recommends research and development in the area of monitoring a diverse set of UTM sensor outputs for detection and identification of anomalous sensor behavior as a technique to enable In-Time System-wide Safety Assurance (ISSA). In addition, research will be performed to develop an in-time methodology for translating out-of-spec sensor data to overall UTM system safety.
CAL Analytics has teamed with Hidden Level to accomplish the following Phase 1 research tasks:
Perform market study of UTM Sensors
Perform study to determine minimum required sensor data elements to enable ISSA
Develop a framework for comparing sensor outputs
Perform study to assess separability of anomalous sensor behavior
Develop requirements and performance metrics for guiding Phase II algorithm development
Perform study exploring methods to assess system safety given degraded sensor performance
Document results of studies in final report with recommendations for Phase II.
The results of this research will form the basis of a UTM ISSA technique that will:
Enable detection and reporting of anomalous sensor behavior through data mining
Identify impacts to UTM system safety using anomalous sensor behavior to assess the impact to overall UTM system safety.
Monitoring anomalous sensor performance within a UTM ecosystem will help NASA research as UTM is transitioned from a research project to productization and commercialization. As NASA shifts gears to UAM, a mechanism to identify, report, and assess the impact of anomalous sensor behavior must exist. This research will also assist with standards development through RTCA and ASTM, by providing insight into standardizing interfaces and reactions to common off-nominal scenarios.
The Non-NASA applications for this project include incorporating the anomalous sensor behavior detection and safety algorithms into a Software as a Service platform that monitors the health and integrity of a UTM ecosystem. This software will automate contingency management to improve UTM system fault tolerance, and ensure safety is maintained, and help automate failure root cause analysis.