ESA Technology Transfer Success Story - Uncovering the secrets of data: spacecraft-grade AI to better inform European SMEs [Aug/2022]
KETTY, a machine learning solution developed by Italian firm
SATE with ESA funding to detect anomalies in satellite telemetry data, has been
used to improve the performance of Austrian IT specialists eMentalist’s event
detection platform for SMEs.
To develop their own system for detecting abnormalities in
satellites, in 2010 ESA contracted engineering firm, SATE, who at the time were
specialised in automobile diagnostics, to develop a solution. Through this
approach SATE developed KETTY (Knowledge ExtracTion Tool for TelemetrY), a software
tool capable of detecting novel behaviour (e.g., the first signs of a fault),
without requiring any a priori knowledge. KETTY uses machine learning
algorithms, which are adapted and used within SATE’s patented methodology:
‘proprietary machine learning approach’. In summer 2017, the Austrian firm
eMentalist, who were at the time involved in the ESA business incubator
programme, were looking to improve the data analytics performance of their
Event Detection platform. SATE used their KETTY algorithm to support eMentalist
in anomaly detection and correlation, and causality analysis. eMentalist sent
SATE their data and SATE returned useful insights, which eMentalist then
incorporated into services sold to companies.
In general, this type of information is useful for companies
who are consumer focused and hope to gain better insight on customer
decision-making. eMentalist leveraged the insights provided by KETTY to develop
a useful product for small and medium-sized firms to gain a better
understanding of their business environment. This improved decision base should
in turn lead to better, more informed companies, boosting their profitability
and success. Potential applications are manifold; to give one example, a firm
considering running a new product can check if this product matches the market
trend.
A number of recent studies have also demonstrated the
enormous potential of data analytics to increase profitability; it was recently
estimated that poor quality data costs the US economy a huge €2.94tn each year,
demonstrating the potential for increased profitability in this area. Just 12% of
companies use data-driven intelligence to guide key business functions and
corporate strategy, highlighting a large potential market. Whilst large companies are to-date
the main users of highly specialised insights products, eMentalist are now
focusing on delivering reliable insight to small and medium-sized firms for
relatively low prices. KETTY is a sophisticated, highly reliable tool which can
be applied to any sector, without requiring specialist knowledge from field
experts.
The full case study report can be accessed in the restricted
area here
(please log in before).
The infographics can be accessed in the restricted area here (please log in before).
This initiative is led by ESA's Technology Transfer and Patent Management Unit (TTPO) in ESA's Directorate for Commercialisation, Industry & Procurement. The Unit is guiding start-ups, entrepreneurs and European businesses in developing spin-offs for ESA's space technologies. More recent successful transfers can be accessed at: Technology Transfer - Funded Projects. For more information, please contact patent@esa.int.