Sivanagaraja Tatinati

Multistep Prediction of Physiological Tremor Based on Machine Learning for Robotics Assisted Microsurgery

For effective tremor compensation in robotics assisted hand-held device, accurate filtering of tremulous motion is necessary. The time-varying unknown phase delay that arises due to both software (filtering) and hardware (sensors) in these robotics instruments adversely affects the device performance. In this paper, moving window-based least squares support vector machines approach is formulated for multistep prediction of tremor to overcome the time-varying delay. This approach relies on the kernel-learning technique and does not require the knowledge of prediction horizon compared to the existing methods that require the delay to be known as a priori. The proposed method is evaluated through simulations and experiments with the tremor data recorded from surgeons and novice subjects. Comparison with the state-of-the-art techniques highlights the suitability and better performance of the proposed method.

Multistep Prediction of Physiological Tremor for Surgical Robotics Applications

Accurate canceling of physiological tremor is extremely important in robotics-assisted surgical instruments/procedures. The performance of robotics-based hand-held surgical devices degrades in real time due to the presence of phase delay in sensors (hardware) and filtering (software) processes. Effective tremor compensation requires zero-phase lag in filtering process so that the filtered tremor signal can be used to regenerate an opposing motion in real time. Delay as small as 20 ms degrades the performance of human-machine interference. To overcome this phase delay, we employ multistep prediction in this paper. Combined with the existing tremor estimation methods, the procedure improves the overall accuracy by 60% for tremor estimation compared to single-step prediction methods in the presence of phase delay. Experimental results with developed methods for 1-DOF tremor estimation highlight the improvement.

A Hybrid Approach for Short-Term Forecasting of Wind Speed

We propose a hybrid method for forecasting the wind speed. The wind speed data is first decomposed into intrinsic mode functions (IMFs) with empirical mode decomposition. Based on the partial autocorrelation factor of the individual IMFs, adaptive methods are then employed for the prediction of IMFs. Least squares-support vector machines are employed for IMFs with weak correlation factor, and autoregressive model with Kalman filter is employed for IMFs with high correlation factor. Multistep prediction with the proposed hybrid method resulted in improved forecasting. Results with wind speed data show that the proposed method provides better forecasting compared to the existing methods.

Physiological Tremor Estimation With Autoregressive (AR) Model and Kalman Filter for Robotics Applications

This paper focuses on developing a simple and efficient tremor estimation algorithm suitable for real-time applications. Autoregressive model in combination with Kalman filter is employed for tremor estimation in robotics devices. A research is conducted with tremor data recorded from surgeons and novice subjects for model identification and characteristics. Results show that appropriate choice of model parameters improves the estimation accuracy. Comparison with existing tremor estimation methods is performed to analyze the performance. Experimental results for 1-DOF tremor estimation are provided to validate the approach.

Automatic Identification of Systolic Time Intervals in Seismocardiogram.

Continuous and non-invasive monitoring of hemodynamic parameters through unobtrusive wearable sensors can potentially aid in early detection of cardiac abnormalities, and provides a viable solution for long-term follow-up of patients with chronic cardiovascular diseases without disrupting the daily life activities. Electrocardiogram (ECG) and siesmocardiogram (SCG) signals can be readily acquired from light-weight electrodes and accelerometers respectively, which can be employed to derive systolic time intervals (STI). For this purpose, automated and accurate annotation of the relevant peaks in these signals is required, which is challenging due to the inter-subject morphological variability and noise prone nature of SCG signal. In this paper, an approach is proposed to automatically annotate the desired peaks in SCG signal that are related to STI by utilizing the information of peak detected in the sliding template to narrow-down the search for the desired peak in actual SCG signal. Experimental validation of this approach performed in conventional/controlled supine and realistic/challenging seated conditions, containing over 5600 heart beat cycles shows good performance and robustness of the proposed approach in noisy conditions. Automated measurement of STI in wearable configuration can provide a quantified cardiac health index for long-term monitoring of patients, elderly people at risk and health-enthusiasts.

A Quaternion Weighted Fourier Linear Combiner for Modeling Physiological Tremor

Goal: This paper offers a new approach to model physiological tremor aiming at attenuating undesired vibrations of the tip of microsurgical instruments. Method: Several tremor modeling algorithms, such as the weighted Fourier linear combiner (wFLC), have proved effective. They, however, treat the three-dimensional (3-D) tremor signal as three independent 1-D signals in the x-, y-, and z-axes. In addition, the force f by which a surgeon holds the instrument has never been taken into account in modeling. Hence, conventional algorithms are inherently blind to any potential multidimensional couplings. Results: We first show that there exists statistically significant subject-and task-dependent coherence between data in the x-, y-, z-, and f-axes. We hypothesize that a filter that models the tremor in 4-D (x, y, z, and f) yields a more accurate model of tremor. We, therefore, developed a quaternion version of the wFLC algorithm and termed it QwFLC. We tested the proposed QwFLC algorithm with real physiological tremor data that were recorded from five novice subjects and four experienced microsurgeons. Although compared to wFLC, QwFLC requires six times larger computational resources, we showed that QwFLC can improve the modeling by up to 67% and that the improvement is proportional to the total coherence between the tremor in xyz and the force signal. Conclusion: By taking into account interactions of the 3-D tremor and the force data, the tremor modeling performance enhances significantly.

Online LS-SVM based multi-step prediction of physiological tremor for surgical robotics

Performance of robotics based hand-held surgical devices in real-time is mainly dependent on accurate filtering of physiological tremor. The presence of phase delay in sensors (hardware) and filtering (software) processes affects the cancellation accuracy. This paper focuses on developing an estimation algorithm to improve the estimation accuracy in the presence of phase delay for real-time implementations. Moving window based online training approach for least squares-support vector machines (LSSVM) is employed in this paper for tremor estimation. A study is conducted with tremor data recorded from the subjects to analyze the suitability of proposed approach for both single-step and multi-step prediction.

Modeling 3D tremor signals with a quaternion weighted Fourier Linear Combiner

Physiological tremor is an involuntary and rhythmic movement of the body specially the hands. The vibrations in hand-held surgical instruments caused by physiological tremor can cause unacceptable imprecision in microsurgery. To rectify this problem, many adaptive filtering-based methods have been developed to model the tremor to remove it from the tip of microsurgery devices. The existing tremor modeling algorithms such as the weighted Fourier Linear Combiner (wFLC) algorithm and its extensions operate on the x, y, and z dimensions of the tremor signals independently. These algorithms are blind to the dynamic couplings between the three dimensions. We hypothesized that a system that takes these coupling information into account can model the tremor with more accuracy compared to the existing methods. Tremor data was recorded from five novice subjects and modeled with a novel quaternion weighted Fourier Linear Combiner (QwFLC). We compared the modeling performance of the proposed QwFLC with that of the conventional wFLC algorithm. Results showed that QwFLC improves the modeling performance by about 20% at the cost of higher computational complexity.

Real-time prediction of respiratory motion traces for radiotherapy with ensemble learning

In this paper, we introduce a hybrid method for prediction of respiratory motion to overcome the inherent delay in robotic radiosurgery while treating lung tumors. The hybrid method adopts least squares support vector machine (LS-SVM) based ensemble learning approach to exploit the relative advantages of the individual methods local circular motion (LCM) with extended Kalman filter (EKF) and autoregressive moving average (ARMA) model with fading memory Kalman filter (FMKF). The efficiency the proposed hybrid approach was assessed with the real respiratory motion traces of 31 patients while treating with CyberKnifeTM. Results show that the proposed hybrid method improves the prediction accuracy by approximately 10% for prediction horizons of 460 ms compared to the existing methods.

Multidimensional Modeling of Physiological Tremor for Active Compensation in Handheld Surgical Robotics

Precision, robustness, dexterity, and intelligence are the design indices for current generation surgical robotics. To augment the required precision and dexterity into normal microsurgical work-flow, handheld robotic instruments are developed to compensate physiological tremor in real time. The hardware (sensors and actuators) and software (causal linear filters) employed for tremor identification and filtering introduces time-varying unknown phase delay that adversely affects the device performance. The current techniques that focus on three-dimensions (3-D) tip position control involves modeling and canceling the tremor in three axes (x-, y-, and z -axes) separately. Our analysis with the tremor recorded from surgeons and novice subjects shows that there exists significant correlation in tremor across the dimensions. Based on this, a new multidimensional modeling approach based on extreme learning machines is proposed in this paper to correct the phase delay and to accurately model 3-D tremor simultaneously. Proposed method is evaluated through both simulations and experiments. Comparison with the state-of-the art techniques highlight the suitability and better performance of the proposed approach for tremor compensation in handheld surgical robotics.