Tactile sensing and display for robot-assisted minimally invasive surgery: Detecting lumps in soft tissue

An important task in clinical diagnosis and surgery is detecting and localizing hard lumps in the soft tissues of the body. Since these lumps are embedded beneath the tissue surface and typically not visible, they must be found through palpation. During open procedures, lumps can be located with relative ease since the surgeon receives natural tactile (cutaneous) feedback while palpating directly with the fingers. In contrast, lump detection becomes considerably more challenging during robot-assisted minimally invasive surgery (RMIS), because distributed tactile information is not sensed or displayed to the surgeon. Artificial tactile feedback (ATF) systems have been proposed to restore tactile information during RMIS that could improve intra-operative assessment and lump localization.;This dissertation describes several studies aimed to improve and evaluate tactile feedback for RMIS lump detection: First, we evaluate the ability of humans and artificial tactile sensors to detect lumps in a constrained, passive, one-dimensional lump detection task. Using a classifier based on signal detection theory, we show that a tactile sensor can detect lumps at lower average applied pressures than a human finger. ii We show the effects of lump size, depth, tissue compliance, and palpation velocity on human and artificial lump detection thresholds. Second, we identify the neural mechanisms underlying lump detection in soft tissue, and demonstrate that firing rate encodes for lump depth, while spatial spread encodes for lump size. We con-dude that lump detection is based on a spatial population code of the SA1 afferents. Third, we develop a novel air-jet lump display that creates a lump percept by directing pressurized air through an aperture onto the finger. We quantitatively characterize the set of achievable pressure profiles and identify the display control mechanisms. Through psychophysical studies, we determine the just noticeable difference (JND) values for the display control parameters. Fourth, we develop an ATF system for teleoperated lump detection in soft tissue. We develop custom algorithms for lump detection and size estimation and show that the ATF system can reliably detect the presence and size of lumps in soft tissue and that the sizes are discriminable through tactile feedback from the air-jet display. Collectively, this dissertation demonstrates the potential of tactile feedback as applied to RMIS, though the methods described can be generally abstracted to traditional minimally invasive approaches as well.