Decision Support System for Fighter Pilots

Publication: ResearchPh.D. thesis – Annual report year: 2007

Documents

View graph of relations

During a mission over enemy territory a fighter aircraft may be engaged by ground based threats. The pilot can use different measures to avoid the aircraft from being detected by e.g. enemy radar systems. If the enemy detects the aircraft a missile may be fired to seek and destroy the aircraft. Such a missile will almost always be either radar guided or heat seeking. It will be launched from a permanent launch pad, or it will be man portable and small enough to fit in the boot of a car. The probability of a missile being detected by onboard sensors depends on the type of missile. If a missile is detected the pilot may choose to deploy electronic countermeasures to avoid the impact of the missile. The countermeasures to choose depends on e.g. the type of missile and guidance system, distance and direction between the missile and the aircraft, an assessment of the environment hostility, aircraft altitude and airspeed, and the availability of countermeasures. Radar systems, guidance of missiles, and electronic countermeasures are all parts of the electronic warfare domain. A brief description of this domain is given. It contains an introduction to both systems working on-board the aircraft and countermeasures that can be applied to mitigate threats. This work is concerned with finding proper evasive actions when a fighter aircraft is engaged by ground based threats. To help the pilot in deciding on these actions a decision support system may be implemented. The environment in which such a system must work is described, as are some general requirements to the design of the system. Decisions suggested by the system are based on information acquired from different sources. The process of providing information from sources such as intelligence, on-board sensor systems, and tactical data from other platforms (aircraft, ships, etc.) is described. Different approaches to finding the combination of countermeasures and manoeuvres improving the pilots survivability is investigated. During training a fighter pilot will learn a set of rules to follow when threat occurs. For the pilot these rules will be formulated in natural language. An expert system can be build by translating these rules into a language understandable by a computer program. This is done in the development of a Prolog based decision support system. A decision support system will base its decisions on input from non-perfect sources. Warnings from on-board sensor can be false and intelligence reports deficient. A Bayesian net is modelled to address this. Building the dependency tables of a Bayesian net requires a large number of cells to be filled with relevant probabilities. Not having sufficient knowledge about these probabilities makes the work with developing a Bayesian net cumbersome. Therefore a method for structural learning is investigated. Here a Bayesian net is build using a set of sample data from a number of missile flight simulations. Knowledge about threats in the current combat scenario may influence the choice of evasive manoeuvres and proper countermeasures. If at any given time more expendables are dispensed than necessary, and none is left for a later necessity, the pilots survivability may decrease. A mathematical model is developed to describe this problem. It is solved to optimality using solver software. When new threats occur the decision support system must be able to provide suggestions within a fraction of a second. Since the time it takes to find an optimal solution to the mathematical model can not comply with this requirement solutions are sought using a metaheuristic.
Original languageEnglish
Publication dateSep 2007
StatePublished
NameIMM-PHD-2007-172
Download as:
Download as PDF
Select render style:
APAAuthorCBEHarvardMLAStandardVancouverShortLong
PDF
Download as HTML
Select render style:
APAAuthorCBEHarvardMLAStandardVancouverShortLong
HTML
Download as Word
Select render style:
APAAuthorCBEHarvardMLAStandardVancouverShortLong
Word

Download statistics

No data available

ID: 4914050