Application of nature inspired optimization algorithms in bioimpedance spectroscopy: simulation and experiment

Accurate extraction of Cole parameters for applications in bioimpedance spectroscopy (BIS) is challenging. Precise estimation of Cole parameters from measured bioimpedance data is crucial, since the physiological state of any biological tissue or body is described in terms of Cole parameters. To extract Cole parameters from measured bioimpedance data, the conventional gradient-based non-linear least square (NLS) optimization algorithm is found to be significantly inaccurate. In this work, we have presented a robust methodology to establish an accurate process to estimate Cole parameters and relaxation time from measured BIS data. Six nature inspired algorithms, along with NLS are implemented and studied. Experiments are conducted to obtain BIS data and analysis of variation (ANOVA) is performed. The Cuckoo Search (CS) algorithm achieved a better fitment result and is also able to extract the Cole parameters most accurately among all the algorithms under consideration. The ANOVA result shows that CS algorithm achieved a higher confidence rate. In addition, the CS algorithm requires less sample size compared to other algorithms for distinguishing the change in physical properties of a biological body.