Simulation of a Detect and Sense System

Abstract

Under current regulations, Unmanned Aerial Vehicles (UAVs) are prohibited from operating in non- segregated airspace. The absence of a pilot is considered to be a potential hazard because the pilot is responsible for the separation with other aircraft. In unmanned aviation the pilot, or operator, can never detect possible separation conflicts without the help of Detect, Sense and Avoid (DSA) systems. To make operation in civil airspace possible, there is a need for certified DSA systems. As of yet there are no certified systems available because the requirements for the certification are not yet developed. To help develop these requirements for these systems, research efforts experiment with these systems in both flight-tests and simulations. One such a research effort is made by the Delft University of Technology (DUT) together with the Netherlands Defence Academy (NLDA). They use a UAV ground station simulator to investigate the ‘Operator-in-the-Loop’ concept: the DSA system does not autonomously take action, but the operator is alerted when a possible threat is detected and asked to verify the threat and to decide what action to take. The research presented in this thesis is performed in order to increase the fidelity of the DSA models used in this simulator. DSA systems use sensors to detect potential threats. This project focused on the simulation of the sensors used for detecting non-cooperative traffic. On the basis of a study of the literature the three most commonly used sensors were selected: radar, electro-optical (EO) and infrared (IR). After the subsystems and interfaces were defined the subsystems were designed in a two step iterative approach. The final radar simulation is a link-budget model and uses the Swerling algorithm for the detection calculation. The EO and the IR simulation system use a simplified geometric aircraft model to calculate the static and dynamic signature on the sensor image. The detection calculation is performed by weighing and combining the signatures and comparing them to a threshold value. All designed systems are verified with data obtained from other research projects.