Optimizing Frequency Selection for Cole-Cole Model Parameter Extraction in Bioimpedance Spectroscopy Using CPE

Wearable bioimpedance spectroscopy (BIS) devices are becoming increasingly popular due to their portability and low power consumption. However, improving power efficiency remains a significant challenge. One proposed solution is to reduce the number of excitation frequencies in the stimulation signal. This study investigates how reducing the number of used frequencies affects bioimpedance spectroscopy performance through simulations using synthetic data derived from a Cole-Cole model incorporating a Constant Phase Element (CPE). Simulations were conducted at various noise levels to evaluate how noise impacts the relationship between frequency and impedance. The results indicate a correlation between the number of frequencies used in BIS and the accuracy of CPE parameter extraction. This is particularly evident for $\alpha$ parameters of 0.8 and higher. Conversely, lowering the $\alpha$ value increases errors, even when a higher number of frequencies is used.