The development of robot-based skill training systems has become a prominent topic
in nursing education, with various innovative robotic systems designed to simulate
real patients, offering a safe and self-directed learning platform for nursing students
to acquire and practice transfer skills. However, many existing patient transfer skill
training systems utilizing humanoid robots are limited in providing high fidelity force
feedback to trainees, particularly in simulating the active movements of patients
during force interactions. The absence of a comprehensive patient transfer motion
model and corresponding control strategy further restricts the effectiveness of such
systems in reproducing real patient behavior. To overcome these challenges, this
study aims to leverage the advantages of manipulators for compliant control and
exploit virtual reality to create an immersive patient transfer skill training system.
In this study, a variable admittance control model was employed to simulate the entire patient transfer motion, revealing a significant positive correlation between the force exerted by nurses and the resulting patient movement. During this phase, a simplified force model was introduced to estimate the force applied by nurses during the transfer process. Subsequently, based on the obtained fitting parameters from admittance modeling, the proposed model was implemented on the manipulator, and a robot system was constructed to replicate the motion of the human waist and its interactions during force-induced maneuvers. To provide a high level of visual accuracy, an immersive virtual patient system was developed by constructing a digital twin model of the robot system. This integration of the robot system with the virtual patient allows the proposed training system to offer both high-fidelity force feedback and visual representations.
In the experimental phase, twelve nursing participants were invited to evaluate the effectiveness of the proposed system. In the initial real human data collection experiment, sensors were employed to measure the actual motion and force applied by nurses. Subsequently, the nursing participants were instructed to perform patient transfers using the proposed robot patient. Objective parameters during the transfers were documented and subjected to comparison. Simultaneously, subjective feedback and evaluations provided by the participants were recorded to assess the reproducibility of the robotic patient in comparison to actual human subjects.
The experimental results obtained from the comparison of objective parameters and subjective evaluations demonstrate that the proposed system, along with the introduced control methods, can effectively replicate the complete patient transfer motion and deliver satisfactory force feedback during the complete transfer. Furthermore, strong support for the practical implementation of the proposed system as an instructional tool was expressed by the participants. These outcomes affirm the effectiveness of the proposed training system and its potential to enhance nursing skills in the future.