Dynamic Virtual Fixture Generation Based on Intra-Operative 3D Image Feedback in Robot-Assisted Minimally Invasive Thoracic Surgery

This paper proposes a method for generating dynamic virtual fixtures with real-time 3D image feedback to facilitate human-robot collaboration in medical robotics. Seamless shared control in a dynamic environment, like that of a surgical field, remains challenging despite extensive research on collaborative control and planning. To address this problem, our method dynamically creates virtual fixtures to guide the manipulation of a trocar-placing robot arm using the force field generated by point cloud data from an RGB-D camera. Additionally, the "view scope" concept selectively determines the region for computational points, thereby reducing computational load. In a phantom experiment for robot-assisted port incision in minimally invasive thoracic surgery, our method demonstrates substantially improved accuracy for port placement, reducing error and completion time by 50% (p=1.06×10-2) and 35% (p=3.23×10-2), respectively. These results suggest that our proposed approach is promising in improving surgical human-robot collaboration.

A Respiratory Motion Prediction Method Based on LSTM-AE with Attention Mechanism for Spine Surgery

Respiratory motion-induced vertebral movements can adversely impact intraoperative spine surgery, resulting in inaccurate positional information of the target region and unexpected damage during the operation. In this paper, we propose a novel deep learning architecture for respiratory motion prediction, which can adapt to different patients. The proposed method utilizes an LSTM-AE with attention mechanism network that can be trained using few-shot datasets during operation. To ensure real-time performance, a dimension reduction method based on the respiration-induced physical movement of spine vertebral bodies is introduced. The experiment collected data from prone-positioned patients under general anaesthesia to validate the prediction accuracy and time efficiency of the LSTM-AE-based motion prediction method. The experimental results demonstrate that the presented method (RMSE: 4.39%) outperforms other methods in terms of accuracy within a learning time of 2 min. The maximum predictive errors under the latency of 333 ms with respect to the x, y, and z axes of the optical camera system were 0.13, 0.07, and 0.10 mm, respectively, within a motion range of 2 mm.