Xiangyu Wang

Perturbation-based training is essential for improving proprioception and reducing fall risk, yet many robotic platforms emphasize sagittal-plane assistance or apply three-dimensional perturbations via trunk, and configurable multidirectional perturbations directly at the foot-ankle interface during walking are still lacking. In this study, we present a robotic system for proprioceptive training that programmatically delivers plantar perturbations in multiple directions during gait. The system integrates a 10-DoF series-parallel hybrid mechanism, an immersive visual interface, and a suite of task modes to reproduce unstable terrains and generate controlled perturbations. To ensure stable operation, a composite control method combining iterative learning with neural-network disturbance estimation was designed, and reduced the trajectory-tracking errors by over 60% in comparison with PID and sliding mode control. Experiments involving ten healthy participants were conducted under mediolateral perturbations and sEMG-based muscle synergy analysis revealed training-induced neuromuscular adaptations, marked by the progressive activation of key frontal-plane balance muscles - gluteus medius, tensor fasciae latae, and peroneus longus - and a reorganization of task-specific synergies under higher perturbation levels. These findings demonstrate that the elaborated proprioceptive training with controlled perturbations drives neuromuscular adaptation, highlighting the potential of the proposed robotic system as a viable tool for balance rehabilitation.

This paper presents an extended direct linear transform (DLT) method for cone-beam computed tomography (CBCT) calibration with circular trajectory, utilizing a simplified phantom with fiducial markers arranged in a straight line. CBCT is a widely-used imaging modality in both clinical and research applications. The calibration is essential for obtaining precise geometric parameters, which are crucial for the high quality of three-dimensional (3D) reconstructed images. However, the accuracy of most phantom-based CBCT calibration methods is highly dependent on the spatial precision of the fiducial markers. Previous calibration phantoms with 3D fiducial arrangements are specifically fabricated and measured with high-precision standards, which are difficult to achieve. A simple line-beads phantom is cost-effective due to its ease of fabrication and implementation, but it presents new challenges arising from the unknown 3D positions of the fiducial markers. To address these issues, this study proposed an extended DLT to transform the original nonlinear problem into a linear one solvable by the least squares method. Then, the inverse of this transform is developed to recover the geometric parameters of the CBCT system and the spatial positions of fiducial markers from the least squares solution. This calibration framework provides an analytical solution for the full set of geometric parameters of the CBCT system. Both simulation and hardware experimental results demonstrate the effectiveness of this method. Compared to the existing methods, the proposed method demonstrates greater robustness to errors in the extracted centroids of projected fiducial markers, with the resulting calibration achieving more accurate target localization and consistently stable reconstruction quality.

Bronchoscopy is increasingly enhanced by the robotic technologies, which facilitate the examination of the airway and raise both diagnostic and therapeutic capabilities for pulmonary diseases. Accurate localization is essential for these robot-assisted endobronchial interventions. As an effective and low-cost solution, visual localization plays a pivotal role in advancing the automation of bronchoscopic robots. However, conventional visual localization approaches are constrained by the homogeneity and texture insufficiency within the bronchial lumen scene, leading to unreliable feature tracking and significant time-consuming image processing. This paper develops an innovative visual localization scheme for bronchoscopic robots, which can accurately estimate the position of the bronchoscope directly from captured airway lumen images. A novel domain transfer network, grounded in transformer technology, can significantly enhance the similarity between synthetic images generated from computed tomograph (CT) data and real video frames. Subsequently, by taking advantage of both personalized generative images and continuous frame sequences, a one-step direct pose estimation module is designed to recover position from endobronchial movements. The proposed method achieves an absolute trajectory error (ATE) of less than 3.85 mm on a human bronchial phantom, representing a more than 55% improvement over previous methods. To the best of our knowledge, this paper achieves the first generative-data-guaranteed visual localization for robotic bronchoscopy, which has a satisfactory accuracy validated by the fluoroscopy.

The complex interactions between flexible needles and tissues present significant challenges in predicting the needle shape during the puncture procedure. In particular, the accurate prediction of flexible needle shape during insertion into complex multilayer tissues, especially when measurement feedback involves non-Gaussian noise, remains an open problem. In this article, we develop a novel reinforcement learning-based active modeling scheme to predict the deflection of the robotic flexible needle. First, the active modeling scheme is constructed by deriving an extended Kalman filter under the maximum correntropy criterion to enhance insensitivity to non-Gaussian noise. Subsequently, based on this scheme, the reinforcement active modeling (RAM) framework is built by incorporating reinforcement learning to compensate for the modeling residuals. Specifically, the theoretical convergence of the proposed scheme is proved by using the Banach fixed-point theorem, thereby ensuring the reliability of needle shape prediction. Finally, a series of comparative experiments is carried out on a self-built robotic flexible needle. The experimental results demonstrate the superior performance of the proposed deflection predictor. Under non-Gaussian noise conditions, the proposed RAM scheme achieves a generalization prediction error reduction of 46.4% in RMSE and over 76.1% in Var during insertion into unknown multilayer tissue.

Accurate tracking of flexible needles is crucial for the success of robotic-assisted puncture procedures in minimally invasive surgery. This paper presents a novel sensor fusion framework that integrates permanent magnet navigation (PMN) and X-ray imaging to enable robust and precise needle tip tracking in robotic-assisted flexible needle (RFN) systems. A magnetic flexible needle is designed by embedding a permanent magnet in a nitinol spring body, allowing wireless localization while supporting the continuous rotation required by duty-cycling steering. To address challenges such as ferromagnetic interference in PMN and occlusions in X-ray, we develop a decentralized extended Kalman filter-based fusion algorithm that dynamically adjusts sensor weights based on cross-covariance, ensuring reliable estimation even with partial sensor failure. Experimental validation on a self-developed RFN platform demonstrates that the proposed framework outperforms single-modality systems under both magnetic interference and visual occlusion conditions, showing improved robustness and tracking continuity. This approach offers a promising solution for image-guided interventions in complex anatomical environments, with the potential to reduce the number of puncture attempts and minimize patient discomfort.

Robotic flexible endoscope (RFE) systems are widely preferred in natural orifice transluminal endoscopic surgeries owing to their unique advantages. Nevertheless, some inherent characteristics of RFE systems (e.g., complex dynamics, input hysteresis, etc.) are still challenging issues to be solved for the autonomous steering. Therefore, this article designs a novel adaptive output feedback control method to handle the unknown dynamics and actuator uncertainties for the bending motion of RFE systems. In particular, a group of continuous signals are first constructed to recover unmeasurable velocities without differential calculation. Then, relying solely on the accessible bending angle, an output feedback controller is proposed based on an adaptive fuzzy logic system, which can simultaneously approximate the uncertain model and input uncertainties online. The closed-loop system is rigorously proven to exhibit exponential stability based on Lyapunov theory. Finally, the proposed approach is applied on a self-built robotic ureteroscopy prototype to validate its effectiveness and robustness. Compared with traditional approaches, the proposed controller has higher tracking accuracy, stronger robustness, and lower sensing requirements. To the best of the authors' knowledge, this is the first output feedback control strategy to achieve accurate autonomous tracking and estimate the unmeasurable signals for RFE systems with unknown dynamics and input hysteresis/deadzone.

In natural orifice transluminal endoscopic surgery (NOTES), the twist steering of the flexible endoscope plays an important role in tracking the preoperative path during cavity intervention. However, the flexible endoscope’s twisting motion has high nonlinearity and uncertainty, which bring challenges for accurate modeling and controller design. In this study, a novel active modeling-based improved control (AMIC) scheme is proposed, which achieves precise control of the robotic flexible endoscope’s twisting motion. First, the Coleman-Hodgdon (C-H) model is modified to serve as the reference model to describe the twist steering. Then, the model error in the C-H model is introduced as an extended state. Upon this, an active modeling algorithm is developed by using the unscented Kalman filter. The proposed model estimates both the twisting angle and the model error in real time. Based on the proposed model, the AMIC strategy is developed to enhance the tracking performance of a proportional-integral-derivative (PID) controller for a reference trajectory. Finally, comparative experiments were conducted on a self-built robotic flexible endoscope under various insertion depths and tip-part configurations. Compared to the PID controller, the experimental results demonstrate that the proposed AMIC scheme achieves a 63.1% reduction in tracking error under a sinusoidal trajectory.