| dc.description.abstract | This thesis is concerned with research related to diabetes treatment, and more specifically with treatment methods via artificial pancreas. A fault detection system is developed and implemented in the APT-simulator. The presented system is classification based, and utilizes SPE monitoring charts. Both a personal and global models are achieved. It is intended that the system is capable of detecting several types of faults, and the presented system accomplishes this by detecting up to 4 different types of faults. In addition to being capable of detecting the different types of faults, the presented system is also able to differentiate between the faults. The fault detection method is tested in the APT-simulator, and is capable of detecting the following faults: loss of amplitude, pressure induced sensor attenuation (PISA), infusion faults, and jumping signal. Furthermore, the method has an acceptable low number of false positives. | |
