Improvements on Non-Invasive Blood Glucose Biosensors Using Wavelets for Quick Faults Detection

G. CAMPETELLI; D. ZUMOFFEN; M. BASUALDO

Journal of Sensors

Hindawi Publishing Corporation

Lugar: New York; Año: 2011 vol. 2011 p. 1 - 11

1687-7268

Non-Invasive Blood Glucose Sensors are still under development stage considering that they are far from being apt for the use in a closed-loop environment. So, every little improvement of these devices is always welcome specially thinking on its important role as a part of an articial pancreas. It was found that some common failures on these sensors are signal shifts and unreal picks that must be taken into account, specially for the importance of computing the correct insulin dosage for diabetic persons. In this context, diabetes technology has been focused since three decades ago on developing several closed-loop control algorithms linking glucose measurements and insulin delivery. Hence, in this paper, a fault detection system based on Discrete Wavelets Transform (DWT) is applied for testing if an erroneous signal from the non-invasive biosensor occurs. The main idea here is to do a correct measurement compensation in order to evaluate the proper insulin dosage through the use of the Predictive Functional Controller (PFC) algorithm. The preliminary study is done by reproducing the faults mentioned above on the blood glucose measurements obtained from a mathematical model of the endocrine system approved by the FDA in 2008 as a substitute to animal trial and recognized as an in silico patient. Then, the simulation environment includes an adult patient with faulty blood glucose measurements and the Fault Diagnosis and Identication (FDI) system, based on DWT, which gives to the PFC algorithm the correct information turning it into a Fault Tolerant Controller (FTC). The nal result is to deliver to the patient the correct insulin dosage.