A dual-mode tactile hardness sensor for intraoperative tumor detection and tactile imaging in robot-assisted minimally invasive surgery

Abstract

Intraoperative tumor detection and shape identification through manual palpation are routinely performed in traditional open surgeries to ensure complete tumor resection. However, most existing robot-assisted minimally invasive surgery (RMIS) systems lack tactile feedback and rely on vision heavily. Traditional tactile sensing methods require the sensor to be placed normal to the tissue surface. But this requirement cannot always be met due to the limited degrees of freedom and the complexity of the environment in confined spaces. This paper proposes a miniaturized piezoelectric tactile sensor for tissue hardness detection by measuring its electrical impedance spectrum. It has two unique detection modes in two orthogonal directions—transverse and longitudinal, and can detect hardness even when the contact angle is large. It is verified by simulations and experiments that both detection modes can be used to detect hardness in the normal contact condition. However, in the case of hardness detection at a large contact angle, the sensitivity of the sensor in the transverse detection mode is significantly higher than that in the longitudinal mode, implying that this mode is more suitable for the large-angle detection. The sensor is then tested on silicone phantoms with hard inclusions and also on an ex vivo porcine liver. In addition, a tactile imaging algorithm based on Gaussian process regression is used to generate the complete hardness distribution of the test sample, which is further processed to extract the shape and boundary of the hard inclusion. The results show that the accuracy of shape detection is high (recall ⩾ 95%, specificity ⩾ 97%), and the smallest feature size it could detect is 1.3 mm. This proves that the proposed tactile sensor has the potential to perform high-accuracy tumor detection in RMIS.