Master Thesis Internship: Big Data for Flight Inspection

It is a misty morning at Schiphol and you are almost back home. “This is your captain speaking” . Looking out the window you can barely see the wingtips. “Please fasten your seatbelt” . Only seconds before touchdown the airport appears from the grey haze. Taxiing to the terminal you wonder if the pilot has super-vision, or if there is something more? Flight inspection Aircraft still rely on a network of ground-based systems for localisation due to their accuracy and reliability. The localisation system used during landing is called the Instrument Landing System (ILS) , which has centimetre accuracy and allows for all-weather landings. To ensure adherence to strict regulations these systems need to be regularly inspected. Currently such flight inspection is performed by NLR using our Cessna Citation II (as can be seen on the picture above). With secondary accurate localisation equipment onboard a prescribed pattern is flown to determine the alignment, sensitivity and strength of the ILS signal. Such flights are expensive and noisy, emit CO2, reduce airport capacity, and with your help, might be unnecessary… Big data Every day thousands of aircraft arrive in the Netherlands, making use of ILS systems. These aircraft all measure the ILS signals as well as their own location, but for each individual aircraft, this data is not accurate enough for flight inspection. However, with the rise of big data, taking advantage of advances in regression methods such as gaussian process regression (GPR), one can combine the data of all these flights to improve the measurement accuracy and use this for flight inspection. This significantly reduces costs and emissions of flight inspection, and allows for increased airport capacity. What will you be doing? The goal of the internship is to develop and test a GPR-based method to use data recorded by regular aircraft to perform (part of the) flight inspection of ILS systems. The assignment consists of: Development of a GPR-based algorithm