Harmonious human-robot collaboration requires the robot to behave like a human partner, which raises the critical question of what factors make the robot do so. This paper proposes a series of policies based on empathetic and non-empathetic intent inference, proactive and reactive action planning, and ego and non-ego action styles to examine which modules enable robots to exhibit human-like behaviors. Two series of experiments are conducted with human subjects to test the performance of the proposed controllers. In Experiment 1, the participant must identify whether the collaborating partner is a human, similar to a Turing test. The classification results empirically verify that the designed empathetic proactive policies enable the robot to exhibit human-like behaviors. Experiment 2 indicates that the proposed policy can be applied to complex collaborative tasks, and this result is consistent with the findings of Experiment 1. From empirical evidence from the experiments, we believe that empathy and proactive policies are essential elements to enable robots to perform human-like actions.

In minimally invasive endovascular interventional surgery, guidewire navigation is an indispensable process. However, even experienced physicians often encounter difficulties in manually manipulating the guidewire for branch selection, while also facing the risk of radiation exposure. In this study, we investigated robotic autonomous guidewire navigation methods. An electromagnetic system was used to track the real-time position and orientation of the guidewire tip, and a state space representing the guidewire within the vascular environment was constructed to guide the robot in precise guidewire manipulation. Experimental results demonstrated that the proposed trial-and-error and centerline-guided methods successfully completed navigation tasks in a static environment, outperforming human navigation performance in terms of trajectory smoothness, trajectory length, and incorrect branch entry counts. For dynamic environment navigation, dynamic time warping (DTW), a technique for measuring the similarity between two temporal sequences, was integrated into the centerline-guided method. The proposed approaches eliminate the need for visual feedback and thereby minimizing the risk of radiation exposure for both patients and medical staff present in the operating room during the procedure.

Detecting concealed cracks before they become reflective cracks is crucial for the proper design and maintenance of roads. However, research on quantitative recognition and intuitive three-dimensional (3D) characterization of concealed cracks is still at an exploratory stage, so this paper proposes a pixel-level segmentation network based on the You Only Look Once for Panoptic (YOLOP) enhancement, which can quantitatively segment the distresses while performing pixel-level detection of concealed cracks in asphalt pavements. Firstly, gprMax was used to numerically simulate and summarize the characteristics of internal cracks in ground-penetrating radar (GPR) images. The relationship between the actual area with crack characteristics and the occupied pixel area in GPR images was then established. The improved YOLOP was utilized to enhance the detection speed and accuracy of concealed cracks, and in terms of a single index, the R-index increased the most by 40%, the P-index increased the least by 26% and the F1-score increased by 37%. Additionally, the segmented crack region was reconstructed in three dimensions using the computed tomography (CT) imaging principle. The vertical height, longitudinal length, cross-sectional area, and volume of the crack region were estimated. This thesis presents important research on the accurate detection and intuitive characterization of concealed cracks in asphalt pavements. It can serve as a theoretical guide for practical engineering applications.

Detecting buried objects using ground-penetrating radar (GPR) profiles typically requires manual interaction and considerable time. To resolve this problem, first, this research used gprMax to model the void defect more realistically and study its GPR image features. The team employed 3D radar to collect 2400 data points of voids from the test road and expressway and used data augmentation to augment the data. Then, convolutional neural network (CNN) algorithms such as Faster R-CNN, RetinaNet, YOLOv3, and YOLOv5, are improved based on the Swin Transformer. In this study, the Swin_YOLOv5 model performed the best with an F1 score of 98.5%, a recall rate of 98.4%, and a precision of 98.7%, and correctly recognized voids while the other three models had false positives (FP). Finally, this research group used a combination of Multiple Screen Shots(MSS) and Swin_YOLOv5_to recognise asphalt pavement voids in real time and achieved an accuracy rate of 89.5% in a field survey to achieve the purpose of real-time detection and identification of voids by GPR.