Data sources in predictive modeling of bird-strike risk

In this article you will find briefing about Anti Bird-Strike analytics and business development within a co-founded airABS

(3 min. read)

Introduction - Tomasz Kowalewski

Being ambitious in solving problems within the highly regulated aviation, require good cooperation with regulatory bodies. In order to build trust, in 2019 we ran a Demo Day during ESTTF European SpaceTech Transfer Forum (see here)

Since then, we took part in number of space-related industry meetings, Horizon2020 Space-Tech Info-day and went through jury assessment up to a finalists level at the Copernicus Masters.

This traction allowed our project group reaching out to the best possible ornithology expertise from Polish Academy of Sciences (PAN). Behavioural knowledge is essential in proper data interpretation, its modeling and shall bring even higher level of predictive models quality.

Together, we have enabled cooperation with Civil Aviation Authority – Bird-Strike Committee and are well positioned for future partnerships in 2020.

Part 1 and 2 - Earth Observation and Navigation data services from space

Tomasz Kowalewski reviewed hard-facts data elements, gathered at obligatory reporting of bird-strike events in aviation, then let the audience understand satellite data sensors and examples from Copernicus Mission. That span from synthetic aperture (SAR), hyperspectral images (CHIME) up to air/water hazards and navigation services used during analysis.

Part 3 - Predictive modeling and analytics

Jacek Stryczyński get the audience through the best practices of predictive data modeling and examples of use of artificial intelligence in image analytics.

He presented selected results of data analysis and conclusions after review of Federal Aviation Administration (FAA) dataset.

Part 4 - DNA sampling and quality assurance

Prof. dr hab. Wies?aw Bogdanowicz (PAN) presented quality conditions necessary to properly identify species using DNA analytics. There is capability to extend assessment results into precise identification using cost-efficient approach.

Part 5 - behavioural ornithology science and models backtesting

Dr hab. Przemys?aw Chylarecki complemented the scientific module by practical examples, showing how much ornithology science helps in contextual data analysis. Only by knowing birds and bats behaviours, data can be narrowed to the most insightful in prediction.

Voluntary questionnaire

Meeting with such a noble group of practitioners in one room was a good opportunity to conduct a survey.

Summary

Limiting the risk of aircraft colliding with animals is in the interests of passengers, cargo shippers, aircraft users and directly saves animals.

The problem is interdisciplinary. Successful modeling require interdisciplinary cooperation. Surely also on an international scale, perhaps under Horizon2020. Once you clearly see an opportunity to join forces, do not hesitate contacting us at airabs.space via an email provided below.

Together with Piotr Paszyński (PJATK) let us warmly thank to prof. dr hab. Tomasz Mazgajski (PAN) for support and Ms Beata Grabowska from Civil Aviation Authority – Bird-Strike Committee for hosting us.

mailto: founders at airabs.space

(c) airabs-earth-observation-for-anti-bird-strike