Go Nakamura

An artificial EMG generation model based on signal-dependent noise and related application to motion classification

This paper proposes an artificial electromyogram (EMG) signal generation model based on signal-dependent noise, which has been ignored in existing methods, by introducing the stochastic construction of the EMG signals. In the proposed model, an EMG signal variance value is first generated from a probability distribution with a shape determined by a commanded muscle force and signal-dependent noise. Artificial EMG signals are then generated from the associated Gaussian distribution with a zero mean and the generated variance. This facilitates representation of artificial EMG signals with signal-dependent noise superimposed according to the muscle activation levels. The frequency characteristics of the EMG signals are also simulated via a shaping filter with parameters determined by an autoregressive model. An estimation method to determine EMG variance distribution using rectified and smoothed EMG signals, thereby allowing model parameter estimation with a small number of samples, is also incorporated in the proposed model. Moreover, the prediction of variance distribution with strong muscle contraction from EMG signals with low muscle contraction and related artificial EMG generation are also described. The results of experiments conducted, in which the reproduction capability of the proposed model was evaluated through comparison with measured EMG signals in terms of amplitude, frequency content, and EMG distribution demonstrate that the proposed model can reproduce the features of measured EMG signals. Further, utilizing the generated EMG signals as training data for a neural network resulted in the classification of upper limb motion with a higher precision than by learning from only measured EMG signals. This indicates that the proposed model is also applicable to motion classification.

A training system for the MyoBock hand in a virtual reality environment

This paper proposes a novel EMG-based MyoBock training system that consistently provides a variety of functions ranging from EMG signal control training to task training. Using the proposed training sytem, a trainee controls a virtual hand (VH) in a 3D virtual reality (VR) environment using EMG signals and position/posture information recorded from the trainee. The trainee can also perform tasks such as holding and moving virtual objects using the system. In the experiments of this study, virtual task training developed with reference to the Box and Block Test (BBT) used to evaluate myoelectric prostheses was conducted with two healthy subjects, who repeatedly performed 10 one-minute tasks involving grasping a ball in one box and transporting it to another. The BBT experiments were also conducted in a real environment before and after the virtual training, with results showing an improvement in the number of tasks successfully completed. It was therefore confirmed that the proposed system could be used for myoelectric prosthesis control training.

A Human Reaching Movement Model for Myoelectric Prosthesis Control

This work was partially supported by a Grant-in-Aid for Young Scientists B Number 26730111.

A Voice Signal-Based Manipulation Method for the Bio-Remote Environment Control System Based on Candidate Word Discriminations

This paper proposes a voice signal-based manipulation method for the Bio-Remote environment control system. The proposed system learns relationships between multiple candidate words’ phonemes extracted by a largevocabulary speaker-independent model and control commands based on a self-learning look-up table. This allows the user to control various devices even if false recognition words are extracted. Experimental results showed that the method accurately discriminate slurred words (average discrimination rate: 93.9±2.40 [%]), and that participants were able to voluntarily control domestic appliances.

A virtual myoelectric prosthesis training system capable of providing instructions on hand operations

This article proposes a virtual hand and a virtual training system for controlling the MyoBock—the most commonly used myoelectric prosthetic hand worldwide. As the virtual hand is controlled using the method also adopted for the MyoBock hand, the proposed system provides upper-limb amputees with operation sensibilities similar to those experienced in MyoBock control. It can also display an additional virtual hand for the provision of instructions on hand operation, such as the recommended posture for object grasping and the trajectory desirable to reach a target. In virtual hand control experiments conducted with an amputee to evaluate the proposed virtual hand’s operability, the subject successfully performed stable opening and closing with high discrimination rates (89.3±6.65%), thanks to the virtual hand’s incorporation of the MyoBock’s operational characteristics. A training experiment using the proposed system was also conducted with eight healthy participants over a period of 5 days. The participants were asked to perform the box and block test using the MyoBock hand in a real environment on the first and final days. The results showed that the number of blocks transported in 1 min significantly increased and that the participants using the instruction virtual hand changed the orientation of the hand approaching blocks from vertical to lateral. The outcomes of the experiment indicate that the proposed system can be used to improve MyoBock hand control operation both quantitatively and qualitatively.

Operation assistance for the Bio-Remote environmental control system using a Bayesian Network-based prediction model.

This paper proposes a Bayesian Network (BN) based prediction model for a layer-based selection and its application to an operation assistance for the environmental control system Bio-Remote (BR). In the proposed method, each node of the BN model is involved in the layer-based selection function, which corresponds to an individual operation command, appliance, etc., and previous logs of operation commands and time division are used as input factors to predict the users intended operation. The prediction results are displayed on the layer-based selection for the BR, and the number of times of operations and time taken for the operations can be reduced with the proposed prediction model. In the experiments, life-logs were collected from a cervical spinal injury patient who used the BR in daily life, and the proposed model was trained based on these recorded life-logs. The prediction accuracy for control devices of the BR system using the proposed model was 84.3 ± 6.5 %. The results indicated that the proposed prediction model could be useful for the operation assistance of the BR system.