Autonomous exploration in unknown environments has become a critical capability of mobile robots. Many methods often suffer from problems such as exploration goal selection based solely on information gain and inefficient tour optimization. Recent reinforcement learning-based methods do not consider full area coverage and the performance of transferring learned policy to new environments cannot be guaranteed. To address these issues, a dual-stage exploration method has been proposed, which combines spatial clustering of possible exploration goals and Traveling Salesman Problem (TSP) based tour planning on both local and global scales, aiming for efficient full-area exploration in highly convoluted environments. Our method involves two stages: exploration and relocation. During the exploration stage, we introduce to generate local navigation goal candidates straight from clusters of all possible local exploration goals. The local navigation goal is determined through tour planning, utilizing the TSP framework. Moreover, during the relocation stage, we suggest clustering all possible global exploration goals and applying TSP-based tour planning to efficiently direct the robot toward previously detected but yet-to-be-explored areas. The proposed method is validated in various challenging simulated and real-world environments. Experimental results demonstrate its effectiveness and efficiency. Videos and code are available at https://github.com/JiatongBao/exploration.

This paper investigates algorithms for enabling surgical slave robots to autonomously explore shape and stiffness of surgical fields. The paper addresses methods for estimating shape and impedance parameters of tissue and methods for autonomously exploring perceived impedance during tool interaction inside a tissue cleft. A hybrid force-motion controller and a cycloidal motion path are proposed to address shape exploration. An adaptive exploration algorithm for segmentation of surface features and a predictor-corrector algorithm for exploration of deep features are introduced based on discrete impedance estimates. These estimates are derived from localized excitation of tissue coupled with simultaneous force measurements. Shape estimation is validated in ex-vivo bovine tissue and attains surface estimation errors of less than 2.5 mm with force sensing resolutions achievable with current technologies in minimally invasive surgical robots. The effect of scan patterns on the accuracy of the shape estimate is demonstrated by comparing the shape estimate of a Cartesian raster scan with overlapping cycloid scan pattern. It is shown that the latter pattern filters the shape estimation bias due to frictional drag forces. Surface impedance exploration is validated to successfully segment compliant environments on flexible inorganic models. Simulations and experiments show that the adaptive search algorithm reduces overall time requirements relative to the complexity of the underlying structures. Finally, autonomous exploration of deep features is demonstrated in an inorganic model and ex-vivo bovine tissue. It is shown that estimates of least constraint based on singular value decomposition of locally estimated tissue stiffness can generate motion to accurately follow a tissue cleft with a predictor-corrector algorithm employing alternating steps of position and admittance control. We believe that these results demonstrate the potential of these algorithms for enabling “smart” surgical devices capable of autonomous execution of intraoperative surgical plans.