Machine learning approach to needle insertion site identification for spinal anesthesia in obese patients.

BACKGROUND: Ultrasonography for neuraxial anesthesia is increasingly being used to identify spinal structures and the identification of correct point of needle insertion to improve procedural success, in particular in obesity. We developed an ultrasound-guided automated spinal landmark identification program to assist anesthetists on spinal needle insertion point with a graphical user interface for spinal anesthesia. METHODS: Forty-eight obese patients requiring spinal anesthesia for Cesarean section were recruited in this prospective cohort study. We utilized a developed machine learning algorithm to determine the needle insertion point using automated spinal landmark ultrasound imaging of the lumbar spine identifying the L3/4 interspinous space (longitudinal view) and the posterior complex of dura mater (transverse view). The demographic and clinical characteristics were also recorded. RESULTS: The first attempt success rate for spinal anesthesia was 79.1% (38/48) (95%CI 65.0 - 89.5%), followed by successful second attempt of 12.5% (6/48), third attempt of 4.2% (2/48) and 4th attempt (4.2% or 2/48). The scanning duration of L3/4 interspinous space and the posterior complex were 21.0 [IQR: 17.0, 32.0] secs and 11.0 [IQR: 5.0, 22.0] secs respectively. There is good correlation between the program recorded depth of the skin to posterior complex and clinician measured depth (r = 0.915). CONCLUSIONS: The automated spinal landmark identification program is able to provide assistance to needle insertion point identification in obese patients. There is good correlation between program recorded and clinician measured depth of the skin to posterior complex of dura mater. Future research may involve imaging algorithm improvement to assist with needle insertion guidance during neuraxial anesthesia. TRIAL REGISTRATION: This study was registered on clinicaltrials.gov registry ( NCT03687411 ) on 22 Aug 2018.

A novel approach to neuraxial anesthesia: application of an automated ultrasound spinal landmark identification.

BACKGROUND: Neuraxial procedures are commonly performed for therapeutic and diagnostic indications. Currently, they are typically performed via palpation-guided surface landmark. We devised a novel intelligent image processing system that identifies spinal landmarks using ultrasound images. Our primary aim was to evaluate the first attempt success rate of spinal anesthesia using landmarks obtained from the automated spinal landmark identification technique. METHODS: In this prospective cohort study, we recruited 100 patients who required spinal anesthesia for surgical procedures. The video from ultrasound scan image of the L3/4 interspinous space in the longitudinal view and the posterior complex in the transverse view were recorded. The demographic and clinical characteristics were collected and analyzed based on the success rates of the spinal insertion. RESULTS: Success rate (95%CI) for dural puncture at first attempt was 92.0% (85.0-95.9%). Median time to detection of posterior complex was 45.0 [IQR: 21.9, 77.3] secs. There is good correlation observed between the program-recorded depth and the clinician-measured depth to the posterior complex (r = 0.94). CONCLUSIONS: The high success rate and short time taken to obtain the surface landmark with this novel automated ultrasound guided technique could be useful to clinicians to utilise ultrasound guided neuraxial techniques with confidence to identify the anatomical landmarks on the ultrasound scans. Future research would be to define the use in more complex patients during the administration of neuraxial blocks. TRIAL REGISTRATION: This study was retrospectively registered on clinicaltrials.gov registry ( NCT03535155 ) on 24 May 2018.

Repetitive control approach towards automatic tuning of Smith predictor controllers.

This paper proposes a new method for automatic tuning of the Smith predictor controller based on a Repetitive Control (RC) approach. The method requires the input of a periodic reference signal which can be derived from a relay feedback experiment. A modified repetitive control scheme repetitively changes the control signal to achieve tracking error convergence. Once a satisfactory performance is achieved through the learning control, the parameters of the Smith predictor controller can be computed from the signals using a nonlinear least squares algorithm. The same relay feedback experiment can provide an initial parameter vector for an efficient implementation of the parameter estimation. Simulations and experimental results will be furnished to illustrate the effectiveness of the proposed tuning method.

Adaptive sliding mode enhanced disturbance observer-based control of surgical device.

This paper presents an enhanced adaptive robust disturbance observer-based motion tracking control methodology. This control approach is established and investigated for a semi-automated hand-held ear surgical device for the treatment of Otitis Media with Effusion. The proposed control methodology is utilised for tracking a desired motion trajectory in the presence of unknown or uncertain system parameters, nonlinearities including hysteresis, and disturbances in the motion system. The stability of the control approach is analysed. The convergence of position and velocity tracking errors is proven theoretically. A precise tracking performance following desired motion trajectory is demonstrated in the experimental study.

Enhanced sensitivity shaping by data-based tuning of disturbance observer with non-binomial filter.

This work presents a new method of sensitivity shaping for motion control systems with disturbance observer (DOB). Although the traditional DOB design has been proved to be an effective means of rejecting disturbances in many applications, shaping of the closed-loop sensitivity is constrained since the Q-filter in DOB is usually restricted to a standard Butterworth or binomial form, where the cutoff frequency is the only tunable parameter. To overcome this limitation and achieve better performance for more demanding motion systems, a general form of the Q-filter is employed in this work. An data-based iteratively tuning procedure is subsequently developed according to a pre-defined cost function, so that the disturbance rejection in the low-frequency band is improved while the high-frequency noise attenuation and robust stability are not compromised. This method is data-based in the sense that the sensitivity shaping process is conducted based on the input-output data collected from the operating motion system, instead of relying on the identified system model. Simulation and experimental results are presented to show the advantage of this data-based sensitivity shaping approach.

Precision Motion Control using Nonlinear Contact Force Model in a Surgical Device.

Otitis media with effusion (OME) is a common middle ear disease in children. The Ventilation Tube Applicator is an ear surgical device to treat OME by inserting a grommet into tympanic membrane (TM). This grommet insertion process requires very delicate control to minimise trauma to the TM and to ensure that the grommet is successfully inserted. This paper introduces a precision motion controller using Hunt-Crossley contact force model to provide a good prediction of the force on the TM during the insertion process. This predicted force was used to reject the disturbance in the motion control scheme and resulted in tremendous improvement in precision and accuracy. The maximum absolute tracking error of the motion controller was reduced by up to 29%, and the root mean square error was reduced by up to 43%. This novel use of a nonlinear contact force model has been proven to work well and can be considered for other robotic surgeries that involves contact and requires placement accuracy.

Assistive technology for ultrasound-guided central venous catheter placement.

This study evaluated the existing technology used to improve the safety and ease of ultrasound-guided central venous catheterization. Electronic database searches were conducted in Scopus, IEEE, Google Patents, and relevant conference databases (SPIE, MICCAI, and IEEE conferences) for related articles on assistive technology for ultrasound-guided central venous catheterization. A total of 89 articles were examined and pointed to several fields that are currently the focus of improvements to ultrasound-guided procedures. These include improving needle visualization, needle guides and localization technology, image processing algorithms to enhance and segment important features within the ultrasound image, robotic assistance using probe-mounted manipulators, and improving procedure ergonomics through in situ projections of important information. Probe-mounted robotic manipulators provide a promising avenue for assistive technology developed for freehand ultrasound-guided percutaneous procedures. However, there is currently a lack of clinical trials to validate the effectiveness of these devices.

Determination of the number of principal directions in a biologically plausible PCA model.

Adaptively determining an appropriate number of principal directions for principal component analysis (PCA) neural networks is an important problem to address when one uses PCA neural networks for online feature extraction. In this letter, inspired from biological neural networks, a single-layer neural network model with lateral connections is proposed which uses an improved generalized Hebbian algorithm (GHA) to address this problem. In the proposed model, the number of principal directions can be adaptively determined to approximate the intrinsic dimensionality of the given data set so that the dimensionality of the data set can be reduced to approach the intrinsic dimensionality to any required precision through the network.

Set-point manipulation approach towards online performance improvement in existing process control loops.

The majority of current industrial process control systems are based on PID control. However, in many of these systems, once the initial setup has been carried out, it is difficult to implement subsequent continuous improvements on the control performance without shutting down the production and disarming the overall system to retrofit alternative controllers. These measures to integrate additional instruments for allowing such flexibility incur heavy costs in terms of time and resources. In this paper, we propose an approach towards achieving the control adaptations which cannot be achieved easily with an existing closed-architectural system. The approach leverages on a set-point manipulation mechanism which allows a virtual modification of the closed-architectural system. In this way, process performance of existing plants can be continuously improved without the need to continuously alter the existing closed loop system. The implementation of the proposed configuration is illustrated with respect to a PID controller although the framework proposed is amenable to higher order controller as well. Simulation examples and experimental results are furnished to show the motivation for such an approach and the improved performance achievable with the proposed approach.

Force estimation and failure detection based on disturbance observer for an ear surgical device.

The disturbance observer is one of the useful tools for estimating the contact force between the subject body and the environment in robotic and mechatronic systems. This paper introduced a novel automatic office-based ear surgical device for the treatment of Otitis Media with Effusion (OME) under the guidance of force sensing information. Since the force sensing information must be reliable so as to ensure the safety of the device, a contact force estimation method based on the disturbance observer is proposed. The system model is built and a control strategy is proposed and developed. In the control strategy, a composite motion controller for an ultrasonic motor (USM) stage is presented, and then the design and the stability analysis of an advanced disturbance observer is given. Furthermore, a contact estimator and a failure detector, aiming to enhance the safety and reliability enhancement, are designed. Finally, the proposed control strategy is studied with both simulation and experiment. The experimental results show that the advanced disturbance observer can estimate the actual contact force correctly and precisely, and the disturbance observer based force estimation and failure detection method is feasible which can be used in force sensing, contact detection and fault diagnosis.

A novel constrained H2 optimization algorithm for mechatronics design in flexure-linked biaxial gantry.

The biaxial gantry is widely used in many industrial processes that require high precision Cartesian motion. The conventional rigid-link version suffers from breaking down of joints if any de-synchronization between the two carriages occurs. To prevent above potential risk, a flexure-linked biaxial gantry is designed to allow a small rotation angle of the cross-arm. Nevertheless, the chattering of control signals and inappropriate design of the flexure joint will possibly induce resonant modes of the end-effector. Thus, in this work, the design requirements in terms of tracking accuracy, biaxial synchronization, and resonant mode suppression are achieved by integrated optimization of the stiffness of flexures and PID controller parameters for a class of point-to-point reference trajectories with same dynamics but different steps. From here, an H2 optimization problem with defined constraints is formulated, and an efficient iterative solver is proposed by hybridizing direct computation of constrained projection gradient and line search of optimal step. Comparative experimental results obtained on the testbed are presented to verify the effectiveness of the proposed method.

Re-visiting the tympanic membrane vicinity as core body temperature measurement site.

Core body temperature (CBT) is an important and commonly used indicator of human health and endurance performance. A rise in baseline CBT can be attributed to an onset of flu, infection or even thermoregulatory failure when it becomes excessive. Sites which have been used for measurement of CBT include the pulmonary artery, the esophagus, the rectum and the tympanic membrane. Among them, the tympanic membrane is an attractive measurement site for CBT due to its unobtrusive nature and ease of measurement facilitated, especially when continuous CBT measurements are needed for monitoring such as during military, occupational and sporting settings. However, to-date, there are still polarizing views on the suitability of tympanic membrane as a CBT site. This paper will revisit a number of key unresolved issues in the literature and also presents, for the first time, a benchmark of the middle ear temperature against temperature measurements from other sites. Results from experiments carried out on human and primate subjects will be presented to draw a fresh set of insights against the backdrop of hypotheses and controversies.

Comprehensive and Practical Vision System for Self-Driving Vehicle Lane-Level Localization.

Vehicle lane-level localization is a fundamental technology in autonomous driving. To achieve accurate and consistent performance, a common approach is to use the LIDAR technology. However, it is expensive and computational demanding, and thus not a practical solution in many situations. This paper proposes a stereovision system, which is of low cost, yet also able to achieve high accuracy and consistency. It integrates a new lane line detection algorithm with other lane marking detectors to effectively identify the correct lane line markings. It also fits multiple road models to improve accuracy. An effective stereo 3D reconstruction method is proposed to estimate vehicle localization. The estimation consistency is further guaranteed by a new particle filter framework, which takes vehicle dynamics into account. Experiment results based on image sequences taken under different visual conditions showed that the proposed system can identify the lane line markings with 98.6% accuracy. The maximum estimation error of the vehicle distance to lane lines is 16 cm in daytime and 26 cm at night, and the maximum estimation error of its moving direction with respect to the road tangent is 0.06 rad in daytime and 0.12 rad at night. Due to its high accuracy and consistency, the proposed system can be implemented in autonomous driving vehicles as a practical solution to vehicle lane-level localization.

Development of an intelligent surgical instrument for otitis media with effusion.

To treat a worldwide common ear disease (OME), a device allowing fast grommet tube insertion has been designed in our earlier works (Gao et al., 2015 [1] and Liang et al., 2013 [2]). However, the instrument has to be manually placed as close as to the Tympanic Membrane before the insertion procedures. To realize a fully automated surgical process, the instrument shall be automatically manipulated to align to the axial direction of ear canal and proceed to complete the surgery. A vision-based servomechanism is proposed to solve the path planning problem. A fuzzy-gain-scheduled controller is proposed to minimize the projection error based on the image detection and the proximity measurement. The proposed controller is proven to outperform the traditional PI controller in pre-clinical trials.

Intelligent vision guide for automatic ventilation grommet insertion into the tympanic membrane.

BACKGROUND: Otitis media with effusion is a worldwide ear disease. The current treatment is to surgically insert a ventilation grommet into the tympanic membrane. A robotic device allowing automatic grommet insertion has been designed in a previous study; however, the part of the membrane where the malleus bone is attached to the inner surface is to be avoided during the insertion process. METHODS: This paper proposes a synergy of optical flow technique and a gradient vector flow active contours algorithm to achieve an online tracking of the malleus under endoscopic vision, to guide the working channel to move efficiently during the surgery. RESULTS: The proposed method shows a more stable and accurate tracking performance than the current tracking methods in preclinical tests. CONCLUSION: With satisfactory tracking results, vision guidance of a suitable insertion spot can be provided to the device to perform the surgery in an automatic way.

Convergence analysis of a deterministic discrete time system of Oja's PCA learning algorithm.

The convergence of Oja's principal component analysis (PCA) learning algorithms is a difficult topic for direct study and analysis. Traditionally, the convergence of these algorithms is indirectly analyzed via certain deterministic continuous time (DCT) systems. Such a method will require the learning rate to converge to zero, which is not a reasonable requirement to impose in many practical applications. Recently, deterministic discrete time (DDT) systems have been proposed instead to indirectly interpret the dynamics of the learning algorithms. Unlike DCT systems, DDT systems allow learning rates to be constant (which can be a nonzero). This paper will provide some important results relating to the convergence of a DDT system of Oja's PCA learning algorithm. It has the following contributions: 1) A number of invariant sets are obtained, based on which we can show that any trajectory starting from a point in the invariant set will remain in the set forever. Thus, the nondivergence of the trajectories is guaranteed. 2) The convergence of the DDT system is analyzed rigorously. It is proven, in the paper, that almost all trajectories of the system starting from points in an invariant set will converge exponentially to the unit eigenvector associated with the largest eigenvalue of the correlation matrix. In addition, exponential convergence rate are obtained, providing useful guidelines for the selection of fast convergence learning rate. 3) Since the trajectories may diverge, the careful choice of initial vectors is an important issue. This paper suggests to use the domain of unit hyper sphere as initial vectors to guarantee convergence. 4) Simulation results will be furnished to illustrate the theoretical results achieved.

Geometrical error compensation of machines with significant random errors.

In this paper, we present a new statistical approach towards soft geometrical error compensation of machines with significant errors. The approach, based on an analysis of the probability of the random error recurring, can reduce the adverse influence of random errors on the compensation of systematic errors. The proposed methodology is made up of three steps. First, error classes are defined from the error bands obtained from calibration. Second, the probability of the magnitude of random error belonging to each of these classes is computed based on the density of the data set within the class. Based on these probabilities, the most probable systematic part of the error measurement can be statistically deduced. Finally, the geometrical error compensation is carried out based on this value. Experimental results are provided for the linear error compensation of a single-axis piezo-ceramic motion system.

Lumbar Ultrasound Image Feature Extraction and Classification with Support Vector Machine.

Needle entry site localization remains a challenge for procedures that involve lumbar puncture, for example, epidural anesthesia. To solve the problem, we have developed an image classification algorithm that can automatically identify the bone/interspinous region for ultrasound images obtained from lumbar spine of pregnant patients in the transverse plane. The proposed algorithm consists of feature extraction, feature selection and machine learning procedures. A set of features, including matching values, positions and the appearance of black pixels within pre-defined windows along the midline, were extracted from the ultrasound images using template matching and midline detection methods. A support vector machine was then used to classify the bone images and interspinous images. The support vector machine model was trained with 1,040 images from 26 pregnant subjects and tested on 800 images from a separate set of 20 pregnant patients. A success rate of 95.0% on training set and 93.2% on test set was achieved with the proposed method. The trained support vector machine model was further tested on 46 off-line collected videos, and successfully identified the proper needle insertion site (interspinous region) in 45 of the cases. Therefore, the proposed method is able to process the ultrasound images of lumbar spine in an automatic manner, so as to facilitate the anesthetists' work of identifying the needle entry site.

Multistability of discrete-time recurrent neural networks with unsaturating piecewise linear activation functions.

This paper studies the multistability of a class of discrete-time recurrent neural networks with unsaturating piecewise linear activation functions. It addresses the nondivergence, global attractivity, and complete stability of the networks. Using the local inhibition, conditions for nondivergence are derived, which not only guarantee nondivergence, but also allow for the existence of multiequilibrium points. Under these nondivergence conditions, global attractive compact sets are obtained. Complete stability is studied via constructing novel energy functions and using the well-known Cauchy Convergence Principle. Examples and simulation results are used to illustrate the theory.

Precision motion control with a high gain disturbance compensator for linear motors.

In this paper, we address the problem relating to the precision control of permanent magnet linear motors to track repeated motion trajectories. A high gain disturbance compensator is developed to improve the control performance degraded due to the presence of significant disturbances. An inverse gain of the overall system model is used to set up a disturbance observer. The observed disturbance is then used to generate a "knocker" signal, to be augmented to the control signal, which can provide the additional energy necessary to overcome the effects of the disturbances. A learning scheme is used to adjust the knocker signal iteratively over the repeated cycles. Simulation and experimental results are furnished to demonstrate the effectiveness of the proposed control scheme.