A novel position measurement approach for a single particle in a channel using Electrical Impedance Spectroscopy (EIS) only with two pairs of electrodes is proposed. The proposed approach is label-free, non-invasive and is very accurate in measurement domain by using machine learning method. Relationship between the single particle’s position and the measured impedances is described by a set of nonlinear equations showing excellent fitting performance with R-square reaching 0.9999. Finding accurate analytical solution of the particle’s position is the inverse problem, it is tackled by a well-trained Support Vector Regression (SVR) model with the help of multi-frequency EIS. The proposed approach is evaluated by simulation models with a single particle in 100 different positions. The results show that the approach performs an outstanding position measurement accuracy to reach 99.25%. Comparing with EIT, it has more simple structure, less measurement time and more accurate measurement results.
- Electrical Impedance Spectroscopy (EIS)
- Multi-frequency measurement
- Particle position measurement
- Inverse problem
- Machine learning