| This dissertation reports the results of performance evaluation of two fault-tolerant systems. The first system is a nonlinear aircraft model---a Generic Transport aircraft Model (GTM), for which the change detection is performed to monitor its health in real-time. The GTM is probed by applying a pulse-compression probing algorithm. The new contribution of this dissertation is to include measurement noise in the system output, and to apply statistical tests to confirm hypotheses on whether there has been a change in the GTM. Aspects of the performance evaluation include detection delay and miss detection. The performance evaluation is demonstrated through the simulation of the GTM. The second system is a location sensor network which is designed to locate an emitter. The emitted signal data are collected by a network of sensor-carrying Unmanned Aerial Vehicles (UAVs). A fault-tolerant guidance criterion is proposed to design the network through optimized way points of UAVs for effective emitter signal acquisition despite the likelihood of sensor loss. The location estimation performance of this system, in terms of the speed and accuracy, is evaluated via the simulation of random experiments. |
