This paper addresses the challenge of safe path planning for mobile robots operating in human-shared environments, where human movements are inherently stochastic. To this end, we propose a reinforcement-learning-based path planning algorithm that accounts for human-related uncertainties at the planning level. The algorithm first employs a Markov decision process learner to explore the environment and generate multiple candidate paths. Second, to reduce computational redundancy, a path eliminator module filters out similar paths based on a proposed diversity metric, ensuring path diversity with minimal overhead. Simultaneously, a Monte Carlo-simulated human risk predictor is integrated into the decision-making unit to select the safest path among the candidates. This integrated algorithm enables robots to generate safe and efficient trajectories without the need for frequent re-planning, even in environments with stochastic human behavior. Simulation results demonstrate the effectiveness of the proposed method. In high-density settings, a $40\times 40$ grid map with 10 humans, the proposed method reduces the average number of conflicts by −69.8%, −54.8%, and −73.4% compared with A*, MDP, and RRT methods, respectively. Meanwhile, it improves task success rates by 94.4%, 70.7%, and 118.75% relative to the same baseline methods.

We proposed a neural network model trained with a small amount of meniscus data (only 144 MR images) to improve the segmentation performance of CNNs, such as U-Net, by overcoming the challenges caused by surrounding tissues. We trained and tested the proposed model on 204 T2-weighted MR images of the knee from 181 patients. The trained model provided excellent segmentation performance for lateral menisci with a mean Dice similarity coefficient of 0.864 (range, 0.743-0.990; SD, ±0.077). The results were superior to those of contemporarily published meniscus segmentation methods based on CNNs.

Various approaches have been studied to solve the path planning problem of a mobile robot designing with multiple tasks. However, safe operation for a mobile robot in dynamic environments remains a challenging problem. This paper focuses on safe path planning for a mobile robot executing multiple tasks in an environment with randomly moving humans. To plan a safe path and achieve high task success rate, a safe planner is developed where a double-layer finite state automaton (FSA)-based risk search (FSARS) method considering environmental risks is proposed. The low-level of FSARS is a novel safe approach to prioritize a safe path rather than merely seeking the shortest path in dynamic environments. Meanwhile, the high-level implements a safety-first search structure utilizing FSA transitions. This structure aims to generating optimal paths while multitasking, avoiding collisions with humans moving completely randomly at the planning level instead of aiming at real-time collision avoidance. FSARS is verified through a series of comparative simulations involving seven types of environmental settings, each with distinct task number, grid size, and human number. We evaluate FSARS based on several metrics, including conflict number, conflict distribution, task success rate, reward, and computational time. Compared with the reinforcement learning method, FSARS reduces the average conflict by 65.4% and improves the task success rate by 34.4%. Simulation results demonstrate the effectiveness of FSARS with the lowest collisions and the highest success rate compared with classic approaches.

We developed an ultracompact three-dimensional (3D) endoscopic scanner that utilizes an electromagnetic-based stereo imaging method (EBSIM) for shape measurements. The method was used to investigate the applicability of imaging fiber optics to featureless surface profiles in confined spaces. The EBSIM is combined with a coherent flexible fiber bundle for structured light pattern projection to ensure that the scanner head can achieve a final diameter of merely 3.4 mm and a baseline length of 1.4 mm. Compared with the existing methods, the scanner is designed to have a minimum focal distance of 2 mm to address the intraoperative evaluation problem. System registration and complex calibration are unnecessary before measurements, and the scanner can be operated from any position and orientation. The scanner is capable of acquiring 3D videos at 30 frames per second. The experimental results demonstrate the feasibility of using the EBSIM for 3D surface profile measurements of featureless cylindrical surfaces. A prototype scanner is presented in extensor, and the average error of a series of known depth distances is found to be equal to 0.15 mm. The implementation shows that the 3D scanner can be potentially applied for endoscopic inspections and intraoperative evaluations in minimally invasive surgeries.