Noninvasive Tissue Identification and Thickness Estimation Using Gap-Changing Capacitive Sensor With Machine Learning Integration

The common goal of surgery is to remove or repair pathological tissues while preserving healthy ones. Minimally invasive surgeries (MISs), such as laparoscopic, endoscopic, or robotic procedures, aim to achieve this with reduced trauma but pose challenges, particularly in tissue identification, due to limited tactile feedback and visual information. Current methods, like intraoperative neurophysiological monitoring (IONM), provide nerve identification but fail to estimate tissue thickness or identify diverse tissues, especially in fluid-filled spaces. To address these limitations, we propose a capacitive sensor that identifies tissue types and estimates thickness based on dielectric constants. The sensor uses interdigitated comb-like electrodes, with adjustable gaps to control the penetration depth of the fringe electric field. In “close-gap mode,” the sensor interacts with the first tissue layer, accurately estimating its dielectric constant ( $\varepsilon _{{1}}$ ). As the gap increases from $50~\mu $ m to 3 mm, the electric field extends deeper, capturing the second layer’s properties ( $\varepsilon _{{2}}$ ). Data collected across multiple gap settings are analyzed via machine learning to classify $\varepsilon _{{2}}$ and estimate the first layer’s thickness ( ${t}_{{1}}$ ) through regression. We validated the sensor through simulations, benchtop tests with materials of varying dielectric constants (\text {1}–2600), and ex-vivo porcine tissue experiments. By utilizing the significant dielectric contrast between different tissues, this sensing approach provides a noninvasive method for identifying tissue layers and measuring thickness. The proposed method has the potential to enhance the precision of tissue removal and minimize nerve damage in MIS.