Desarrollo y evaluación de algoritmos de identificación no cooperativa de blancos radar reales mediante una base de datos sintéticaaproximación hacia una solución operativa.

Laburpena

One of main technology challenges concerning the Security and Defence areas is the rapid and reliable identification of different aircraft flying over the battle space or over civilian facilities with high level of security, such as airports, governmental buildings, etc. The different techniques that have been researched to solve this problem may roughly be divided into two groups: cooperative and non-cooperative techniques. Both are based on radar technology due to its capability of long range day/night operation, as well as smoke, weather and foliage penetration. Cooperative systems are based on the communication between an identification system and the unknown target. The identification system sends an interrogating signal to the target, which replies automatically providing its identification. However, cooperative systems are able to identify just friend or neutral aircraft, and become more vulnerable in some scenarios, such as armed conflict areas where communications are often disrupted by the other side. In principle, this lack of identification could be covered by non-cooperative systems, which do not require any communication with the target. A non-cooperative system sends a signal to the target that is reflected by different parts of the aircraft. All these reflections form a global reflected signal that provides information about the target geometry, which is known as the target signature. Finally, the measured target signature is compared with other target signatures of a reference database. To bring into operation a non-cooperative system is mandatory a proper signal processing of the acquired data, a good classification method and a huge database populated with as much information as possible. Up to now, the operative non-cooperative systems are equipped with a database populated with measurements of actual targets, which is a big drawback because these systems are able to identify just the aircraft which have collaborated in the measurement campaigns previously carried out in order to create this database. This PhD dissertation is aimed to develop a non-cooperative system able to identify targets at any aspect angle in a short period of time. In order to do this, a new identification methodology is presented and demostrated, where different acquisition techniques, types of signal processing and classification methods have been considered and analyzed to achieve that goal. However, the main contribution of this research is related with the fact that the information populating the database is generated by electromagnetic simulation software tools, as it requires lower cost and allows a faster way to obtain information of all the possible targets, aspect angles and configurations.