Significant advancements have been made in applying machine learning techniques to the classification of surface electromyography (sEMG) signals. Despite these improvements, there remains considerable room for enhancing these methods to achieve even greater accuracy. This is particularly evident in complex sEMG classification problems, where the patterns across different classes can be very similar. This similarity often leads to reduced accuracy in both validation and test sets, highlighting the need for more refined approaches. This study aims to improve sEMG-based hand gesture recognition by introducing a novel Stacked One-vs-One (SOvO) classification approach. Traditional One-vs-One (OvO) strategies decompose multi-class problems into binary classifications but often suffer from cumulative errors of weak classifiers. The proposed SOvO method enhances classification by stacking the probability outputs from pairwise binary classifiers in the OvO approach on the training data as additional features for the final classification. Depending on whether the final classifier is a deep learning or conventional model, different feature integration strategies were explored. A comparative analysis was conducted on the Ninapro DB4, DB5, and DB6 datasets across two window sizes. Results show that SOvO consistently outperforms OvO in accuracy by 3% to 10%. This improvement is statistically significant, as confirmed by a repeated measures ANOVA ( p < 0.001). These findings demonstrate that SOvO provides a more robust and accurate framework for sEMG-based gesture recognition, addressing limitations of conventional OvO methods.

The real-time use of electromyography (EMG)-based mechatronic rehabilitation devices aiming to detect stroke patients' hand grasping intention is hindered by a significant concern: the lack of robustness against variations in EMG signal patterns due to arm posture changes. This problem results in degraded EMG signal measurements and inaccurate recognition of muscle patterns. Several studies have aimed at tracking changes in EMG patterns by placing multiple EMG sensors around the forearm and developing a classifier using data collected from various arm postures recorded by all sensors. Although these methods show promise, the significant computational resources required for real-time data processing become notable concerns when using multiple EMG sensors. To address these challenges, this study introduces a novel approach that aims to reduce the number of EMG channels that need to be processed. The study proposes a new optimal-channel-selection technique, coupled with a convolutional neural network (CNN), which selects two out of eight EMG channels within an armband based on the arm posture and individual demographics. As a result of using only two channels rather than the entire array (eight channels), the user's grasping intention prediction time took only 2.3 seconds with a classification accuracy of around 81%. In comparison, the commonly used eight-channel method took 8.6 seconds for grasping intention detection with an accuracy level of 79%. These findings show potential in tackling the challenge of EMG measurement degradation caused by arm motion, offering a path towards enhanced accuracy and quicker responsiveness in EMG-based mechatronic rehabilitation devices.

Wearable devices can suppress or reduce hand tremor motion associated with neurological disorders such as Parkinson's disease. Tendon-driven transmission systems have been proposed as a way to decrease the size and weight of these devices; however, they have complex control system requirements due to their substantially nonlinear behavior. To address this issue, this study focused on the development of a comprehensive kinematic model of a wearable tremor suppression glove that more accurately calculates the tendon displacement during hand motion. The novelty lies in the identification of the threshold bending angle for each joint at which the driven tendon touches the arc of the joint. The derived kinematic model of the glove was verified by both simulation and benchtop experiments, and the proposed model was validated during single-joint and multijoint hand movements. The kinematic model shows a mean 2-D correlation coefficient of 0.96$\boldsymbol{\pm }$0.01 with the experimental data. Compared to the Euclidean norm model presented in the literature, it presents an average 83% improvement (a 4%–96% reduction in root mean square errors depending on the joint), which is most significant for increasing tendon displacements.

In spite of the remarkable benefits that minimally invasive surgery provides for patients, the absence of force feedback is still a significant disadvantage. Several studies have been performed to address this issue; however, an accurate sterilizable force sensing technology for measuring axial and grasping forces is still missing. In this work, an innovative partial grasper has been designed and developed for a laparoscopic needle driver that can measure axial and grasping force information at the grasper tip. Fiber Bragg Grating sensors are used in this work because of their sterilizability and high sensitivity. Accuracies of 0.19 N and 0.26 N were achieved for the grasping and axial sensors respectively.

Modern joint replacements provide an effective treatment for patients suffering from debilitating joint conditions; over 300,000 total hip replacement surgeries are performed each year in North America. Nonetheless, about 10% of all hip implants need to be replaced within 10 years of patients receiving a total hip replacement, typically due to component loosening or infection. The capabilities of state-of-the-art 3D metal printing systems allow research groups and medical device manufacturers to explore entirely new designs for more intelligent and effective orthopedic implants, which may offer improved performance and more effective monitoring. This paper presents the development of a platform for a next generation “smart” implant: an instrumented implant with electronic sensors and wireless telemetry. This implant utilizes an ultra-low power microcontroller and wireless radio frequency transceiver onboard. Additive manufacturing in medical grade alloy has been used to create a modified version of a commercial hip implant with allowance for an electronics cassette, including an instrumentation bay with a 2.4 cm3 volume. This modular approach provides the capability of incorporating a range of sensors in the future (e.g., sensors measuring acceleration, stress, strain, and temperature) during the design, optimization, and testing of next-generation implants.

Tumor localization in traditional lung resection surgery requires manual palpation of the deflated lung through a thoracotomy. It is a painful procedure that is not suitable for many patients. Therefore, a multisensory mechatronic device was designed to localize tumors using a minimally invasive approach. The device is sensorized with tactile, ultrasound and position sensors in order to obtain multimodal data of soft tissue in real time. This paper presents the validation of the efficiency and efficacy of this device via an ex vivo experimental study. Tumor pathology was simulated by embedding iodine-agar phantom tumors of varying shapes and sizes into porcine liver tissue. The device was then used to palpate the tissue to localize and visualize the simulated tumors. Markers were then placed on the location of the tumors and fluoroscopic imaging was performed on the tissue in order to determine the localization accuracy of the device. Our results show that the device localized 87.5% of the tumors with an average deviation from the tumor center of 3.42 mm.

The reduced access conditions present in Minimally Invasive Surgery (MIS) affect the feel of interaction forces between the instruments and the tissue being treated. This loss of haptic information compromises the safety of the procedure and must be overcome through training. Determining the skill level of trainees is critical for ensuring patient safety. The objective of this work was to evaluate the usefulness of force information for skills assessment during MIS. Experiments were performed using a set of sensorized instruments capable of measuring instrument position and tissue interaction forces. The results show that experience level has a strong correlation with force-based metrics. The proposed metrics can be automatically computed, are completely objective, and measure important aspects of performance.

This paper outlines a framework for the preoperative planning of robotics-assisted minimally invasive cardiac surgery with application to coronary artery bypass grafting. The intent of the proposed framework is to improve surgical outcomes by considering the intraoperative requirements of the robotic manipulators and the anatomical geometry of the patient's chest. This includes target reachability, instrument dexterity for critical surgical tasks and collision avoidance. Given the patient's preoperative chest computed tomography images, the planning framework aims to determine the optimal location of the access ports on the ribcage, along with the optimal pose of the robotic arms relative to the patient's anatomy. The proposed multi-objective optimality criteria consist of a measure of clearance as well as a new collective kinematic measure. The minimum distances among the robot arms provides a measure for the likelihood of collisions. The proposed kinematic measure is composed of two modified manipulability indices that are dimensionally homogeneous and, in contrast to previously-used measures, are more likely to yield isotropic force and torque distributions when optimized for surgical interventions. The results of a case study illustrate the compatibility of the framework with general guidelines used by experienced surgeons for port selection. Furthermore, the framework surpasses those guidelines by ensuring the feasibility of the solutions in the sense of collision avoidance and surgical target reachability.

The limitations of minimally invasive surgery include the inability to sense forces exerted by the instruments on tissue and the limited visual cues available through the endoscope. A modular laparoscopic instrument capable of measuring force and position has been designed to address these limitations. Novel image-based position tracking software has been developed and integrated within a graphical user interface. This modular system is low cost, versatile, and could be used for training, localization of critical features or for guidance during surgical procedures.

An integrated system for minimally invasive robot-assisted image-guided lung brachytherapy has been developed. The system incorporates an experimental setup for accurate radioactive seed placement with commercially available dosimetry planning software. The end result is a complete system that allows planning and executing a brachytherapy procedure with increased accuracy. The results of the in vitro seed placement evaluation show that seed misplacement has a significant effect on the volume receiving more than 200% of the dose (V200), and the minimum dosage received by 90% of the volume (D90).

Minimally invasive surgery is a technique that provides numerous benefits to the patient, but presents challenges to the surgeon in that dexterity, hand-eye coordination and haptic perception are compromised. Robot-assisted minimally invasive approaches have addressed the problems of dexterity and coordination; however, the lack of kinesthetic and tactile feedback remains a significant drawback. Despite many advances in this area, little is currently known about what level of feedback performance is adequate to allow the surgeon to palpate tissue to detect an underlying tumour. This paper describes experiments that were conducted on ex-vivo porcine lung, using artificial tumours, to elucidate one measure of sensor performance required to detect the presence of a tumour. The results indicate that a force- sensitive probe with a sensing range of 0 to 10 N and a resolution of 0.01 N would allow a tumour to be localized via palpation using kinesthetic feedback.

An experimental test-bed for robot-assisted image-guided minimally invasive lung brachytherapy has been developed. The setup includes surgical robots, a device to accurately drop brachytherapy seeds, an ultrasound system, electromagnetic trackers, and navigational software. An evaluation using the test-bed was performed to compare the open procedure and the video-assisted thoracoscopic procedure with the proposed robot-assisted minimally invasive procedure under ultrasound image guidance and electromagnetic tracking. The results show that the use of robotic systems and image guidance can provide improved seed placement accuracy as compared to open surgery, while reducing the invasiveness of the procedure and trauma to the patient's body.

A device for robot-assisted brachytherapy has been designed for use with a minimally-invasive surgical robot such as the ZEUStrade It may be loaded with a standard brachytherapy needle allowing the robot to position its tip at a specified location within cancerous tissue. The device accurately retracts a hollow needle loaded with radioactive seeds while a stationary plunger pushes the seeds out into the tissue. The position error when retracting the needle is less than 0.05 mm. The use of this device, together with image-guidance for needle placement, can improve radiation dose delivery when treating certain types of lung cancer. The incorporation of this device into robotic systems that are already approved for clinical use can potentially allow it to be commercialized much sooner than competitive technology

Accurate and proper placement of ports during robotically-assisted endoscopic surgery is critical to the success of the procedure. Improper placement of ports can lead to robot collisions, the inability to reach the surgical site, the inability to manipulate the tools properly, or collisions between the tools inside the patient’s body. In current practice, port placement methods do not consider the ability of the robot to manoeuvre the tools. Furthermore, there is a lack of guidance on what the position of the robot or the configuration of its arms should be. This paper proposes to choose the best port location and determine the position of the base and arms of the robot such that its performance is maximized. Performance measures developed in the past to optimize robot design and manipulation are used to optimize the placement of ports when using a surgical manipulator for artery dissection during coronary bypass surgery.