Data Mining for Astronauts Medical Autonomy

The paper "Data Mining for Astronauts Medical Autonomy", written in collaboration with ESA following the AMMD (Autonomous Medical Monitoring and Diagnostics) project, was presented during the 14th International Conference on Space Operations (SpaceOps 2016).

The European Space Agency (ESA) has recently patented and developed two data mining techniques that support non-medical telemetry monitoring and diagnostic tasks. Novelty Detection allows finding unusual behaviour in data which often is the signature of a developing anomaly. DrMUST supports characterization and anomaly investigation. It finds when similar behaviours have happened in the past and discovers possible causes for a given anomaly. Both Novelty Detection and DrMUST are highly valued by Flight Control Engineers and have become the de-facto standard data analysis tools at the European Space Operations Centre. In this work we evaluate how these ESA developed techniques and other data mining approaches can be of benefit for autonomous medical monitoring and diagnostics with the goal of medical autonomy in mind. Medical autonomy is a precondition to enable long cruise human spaceflight missions. In unclear medical conditions or even medical contingencies, astronauts should have the means to collect their medical data and to understand if their physiological parameters are within nominal rangess, get early warnings if they are not, are provided with possible diagnoses and get practical recommendations and treatment options to efficiently solve the medical issue. The target users group are astronauts and medical teams with possible spin-off to non-space related applications in the medical domain. The scenario will be, for the astronauts and the crew medical officer, to be able to obtain early awareness and resolution proposals of their health status, much faster than if interaction with ground would be needed. The definition, design and validation of data mining algorithms that enable this medical scenario are discussed in this work. In order to prototype these scenarios, anonymised medical data with relevance to the space domain has been used from several sources. The paper describes in detail the data, process, the data mining techniques that are helpful in the context of medical autonomy and the assessment from the ESA medical team. Future work include the spin-off of this project to non-space health care.

Download the full article at the following link
Data Mining for Astronauts Medical Autonomy

Comments are closed.