Abstract
Tissue composition assessment techniques are used to help better comprehend physiological processes and their overall impact on the biological state of the experiments subjects. The research presented in this manuscript aims to design a bioimpedance spectroscopy integrated measurement system capable of measuring a wide range of biomarkers over long periods of time (up to one year). The presented measurement system can be used for physiological variables long time monitoring applications in general. Nevertheless, the presented solutions target in particular fish species in the context of the POPSTAR project which aims to enhance our understanding of fish behavior by analyzing not only the environment in which fish travel and live but also the fish themselves. After identifying the design challenges of a bioimpedance spectroscopy integrated measurement system, we have proposed a novel hybrid architecture providing fast bioimpedance spectroscopy while maximizing the measurement precision. As the signal generation blocks are critical and their performances affect the whole architecture performances. The second part of this research focuses on the design and optimization of the signal generation part of the architecture. Indeed, we have enhanced the stimuli generation signals quality for single tone and multitone excitations while proposing for this blocks low complexity on-chip implementations. In the last part of our work the current driver that transforms the voltage stimuli into an excitation current is discussed. A novel analog topology using an improved regulated cascode and a common-mode feedback compensation independent of process variations is presented. The first chip prototype implementing the critical blocks of the bioimpedance integrated measurement architecture has been designed and simulated in a 0.18 µm AMS (Austria MicroSystems) CMOS process operating at 1.8 V power supply.