Yu Nishiyama

Parameter extraction for identifying product type of mckibben pneumatic artificial muscles

This study demonstrates that various unknown parameters used in nonlinear models of McKibben pneumatic artificial muscles (PAMs) can characterize the features of McKibben PAM products. By focusing on a parameter space in the PAM model, this study employs a support vector machine (SVM) to determine which unknown parameters characterize each PAM product. For validation, three different PAM products are analyzed to observe whether the resulting minimal combination of parameters will help to identify the product. The observation is expected to provide prior PAM knowledge that can be used to develop efficient parameter estimation and capture aging degradation, which are important for robust estimation and control in PAM systems.

Filtering with state-observation examples via kernel monte carlo filter

This letter addresses the problem of filtering with a state-space model. Standard approaches for filtering assume that a probabilistic model for observations (i.e., the observation model) is given explicitly or at least parametrically. We consider a setting where this assumption is not satisfied; we assume that the knowledge of the observation model is provided only by examples of state-observation pairs. This setting is important and appears when state variables are defined as quantities that are very different from the observations. We propose kernel Monte Carlo filter, a novel filtering method that is focused on this setting. Our approach is based on the framework of kernel mean embeddings, which enables nonparametric posterior inference using the state-observation examples. The proposed method represents state distributions as weighted samples, propagates these samples by sampling, estimates the state posteriors by kernel Bayes’ rule, and resamples by kernel herding. In particular, the sampling and resampling procedures are novel in being expressed using kernel mean embeddings, so we theoretically analyze their behaviors. We reveal the following properties, which are similar to those of corresponding procedures in particle methods: the performance of sampling can degrade if the effective sample size of a weighted sample is small, and resampling improves the sampling performance by increasing the effective sample size. We first demonstrate these theoretical findings by synthetic experiments. Then we show the effectiveness of the proposed filter by artificial and real data experiments, which include vision-based mobile robot localization.

An automatic templates selection method for ultrasound guided tumor tracking

In this report, we propose a novel robust tumor tracking method for ultrasound guided RFA (radiofrequency ablation) treatments. Organ deformations seriously deteriorate the tracking performance. To cope with this problem, we propose a novel motion tracking method using dynamic templates. Our method achieves stable tracking by selecting templates automatically based on the texture features of ultrasound diagnostic images. On the other hand, the conventional method is unstable due to the variation to select templates manually. Experimental results show the effectiveness of our proposing motion tracking method concerning the robustness and accuracy.

A study for tracking focal lesions in non-invasive ultrasound theragnostic system

In recent years, HIFU (High Intensity Focused Ultrasound) therapy, which is one of the non-invasive therapies utilizing focused ultrasound, have attracted a great attention as a novel treatment method for a focal lesion, such as a tumor and a stone. However, the focal lesion moves in accordance with respiration, which may cause the damage for the surrounding normal tissues. To cope with this problem, we have developed a non-invasive ultrasound theragnostic system (NIUTS). In this report, we propose a novel tracking method, which is implemented in NIUTS, based on “Partial Active Shape Model” to enhance the servo performance for the focal lesion. Experimental results shows the effectiveness of the proposed servo method concerning the precision and robustness for a kidney phantom.

Efficient PSO-based algorithm for parameter estimation of McKibben PAM model

This study considers the parameter estimation problem for an elaborate nonlinear hybrid model of a McKibben pneumatic artificial muscle (PAM) actuated by a proportional-directional control valve and proposes an efficient particle-swarm-optimization-based algorithm to find adequate model parameters in terms of model accuracy and computation time. A novel approach to making an algorithm more efficient is to focus on the parameter space of the PAM model and to use a support vector machine (SVM) to specify a subset in the parameter space. The inertia of the PSO algorithm is erased to the extent that the particles are allowed to search intensively in the subset region. Furthermore, this study validates the efficiency of the proposed algorithm using three different practical PAM products.

Robust servoing method for renal stones/tumors for the non-invasive ultrasound theragnostic system

The main problem on HIFU (High Intensity Focused Ultrasound) therapy is the difficulty to locate HIFU focus precisely onto the focal lesion, which is located in the moving organ such as livers/kidneys, due to the deformation and rotation, which is caused by respiration. Furthermore, rib bones frequently block the acoustic path to the lesion. The acoustic shadow, which is generated by the rib bone, is observed in the ultrasound images and the lesion is hidden in the shadow. To cope with this problem, we have developed a novel method to track, follow, and monitor the lesion by utilizing the contour information of the organ, which incorporate the lesion under those difficult conditions. As for the tracking method, the contour of the organ, which is deformed and rotated in accordance with respiration, is extracted automatically in 2-D ultrasound images. The missing contour information in the acoustic shadow area is estimated and compensated by the surrounding contours, which are successfully extracted. To con...

Liver tracking system utilizing template matching and energy function in high intensity focused ultrasound/radio frequency ablation therapy

Monitoring and evaluating the therapeutic effects, in ultrasound images during HIFU (High Intensity Focused Ultrasound) and RFA (Radio Frequency Ablation) therapies, are important. However, the common problem is the difficulty to monitor the progress and identify the positions of the focal lesion precisely in accordance with the progress of the treatment. This problem is caused by the movement of organs and the change of the textures information in ultrasound images. To overcome those problems, we have developed a novel method to track, follow, and monitor the focal lesion by the combination of the energy function method and the template matching method. The template matching method uses the characteristic texture information of the organ near the focal lesion. The energy function is implemented in order to reinforce the robustness of the tracking performance of the template matching method. Particularly, we evaluate the existing probability of the focal lesion considering the continuity of the movement a...

An ultrasound guided monitoring system for high intensity focused ultrasound and radio frequency ablation therapies

In accordance with the progress of the HIFU (High Intensity Focused Ultrasound) and RFA (Radio Frequency Ablation) treatments, the intensity of the focal lesion and the surrounding marginal area changes little by little. Here, it should be noted that the human eyes are very weak when the image takes a long time to change even if the image change from the start of the ablation treatment is large. To cope with those problems, we propose an ultrasound guided monitoring system to evaluate the progress of the ablation therapy quantitatively in order to secure the certain level of the monitoring during the HIFU and RFA therapies by reinforcing the lack of experience of medical doctors. Particularly, we propose a system to monitor and evaluate the intensity of the focal lesion and the intensity of the surrounding marginal area by tracking the position of the characteristic texture area near the focal lesion, which moves in accordance with the respiration. Our method to overlay the focal lesion and the surroundin...