Masataka Yoshioka

Power assistance for human elbow motion support using minimal EMG signals with admittance control

In this paper, different from other researches, we developed an admittance control model with EMG signal to control our developed power assistance device for human elbow motion support. A two degree of freedom power assistance device is developed, and a simple but effective EMG signal processing method is proposed. Further, an admittance control model using EMG signal is introduced to generate speed command based on the EMG signal to perform power assist task. It can easily and arbitrarily implement different power assist ratio with different parameters in admittance model. Experimental results show validity of the proposed approach.

Power assistance on slope of an omnidirectional hybrid walker and wheelchair

In this paper, first we briefly present an overview of our developed electric powered user-propelled hybrid walker and wheelchair robot. Then we focus on analysis and investigation of a movement in which the robot starts to ascend and to descend a slope. We originally point out that in such motion the users walking speed and the robots traveling speed are inevitably different since the robot is undergoing a planar motion. This explains why the load of the user suddenly increases/decreases when a user-propelled walker or wheelchair starts to ascend/descend a slope. Note that as far as our knowledge this is the first time clarifying the intrinsic nature of the movement. Further we propose two speed compensation approaches under an admittance based interaction control to let the user get a smooth “load feeling” as the user utilizing the robot on a horizontal plane. The experiments verify the effectiveness of the proposed approaches.

EMG estimation from EEGs for constructing a power assist system

Brain-Machine Interface (BMI), has been becoming gradually an effective way to help and support the disabled person by using his/her brain activity information. It is also expected to be used for both of the disabled and the healthy people. In this paper, aiming to estimate the force/torque information from brain activity to help and support humans daily life, we estimate the humans muscular activity from elec-troencephalography (EEG) by Principal Components Analysis (PCA) and Recursive Least Squares (RLS). The concept of the proposed approach using PCA and RLS is explained, and then the proposed approach is verified by experiments. The results show that the estimation of electromyography (EMG) from EEG is possible and this implies a great potential to use EEG signals for supporting human activities.

Investigation of the EEG scalp distribution for estimation of shoulder joint torque in the upper-limb power assistant system

Brain-Machine Interface (BMI), has been considered as an effective way to help and support the disableds daily lives and rehabilitation to use their brain activity information instead of their bodies. As an extension, it is also expected to be used for healthy individuals. In our study, we try to estimate the necessary force/torque information for a subject from his/her EEG signals when using an exoskeleton robot to perform the power assistance of the upper limb without using external force sensors and EMG sensors. For real time measurement and control of the system, it is necessary to reduce the number of electrodes and signal processing burden. Therefore, it is important to know the EEG distribution on the scalp, and hence extract the motion-related information from multiple measurement locations, effectively and exactly. In this paper, at first, the contribution of each electrode is analyzed by PCA(Principal Component Analysis). Then, ICA (Independent Components Analysis) is used to extract the source information of neural components to obtain the location of each source information by plotting the components at the scalp to determine which measurement location should be measured to obtain effective signals related to motion. At last, the transition of the EEG signals with time is analyzed, and the results are graphically plotted. The results show that the changes of EEG signals, related to the motion of shoulder joint, are distributed throughout the brain cortex, and the locations of electrodes at C3, C4, Cz, P3, P4 are the best ones for obtaining useful and effective information of motion consciousness. This discovery brings great benefit in the use of EEG signals for supporting activity for both healthy and disabled individuals by controlling a brain-machine interface.

Power assistance of an omnidirectional hybrid walker and wheelchair with admittance model and Iterative Learning Control

In this paper, at first we briefly present an overview of our developed hybrid walker and wheelchair robot. Then we focus on analysis and investigation of the movement in which the robot starts to ascend/descend a slope, and point out that the users walking velocity and the robots velocity are inevitably different. Further, in order to let the user get a smooth load feeling as the user use the robot on a horizontal plane, we propose V.C (Velocity Compensation) and ILC (Iterative Learning Control) to adaptively adjust the control parameter. The experiments verify the effectiveness of the proposed approaches.

Motion support of upper extremity with agonist alone under negative admittance control

In this paper, different from the conventional admittance control, an originally negative admittance control, in which two virtual parameters, virtual damping coefficient and virtual moment of inertia, are negative, is developed to support the movement of humans upper extremity only by agonists EMG signals. First, we empirically determine two agonists, biceps brachii and clavicular part of deltoid to get their EMGs respectively as the control signals for elbow joint and shoulder joint. Then a negative admittance control is developed to support the movement of humans upper extremity for holding up, keeping and holding down a load only by agonists EMG signals. The experiments demonstrate that seamless movements of holding up, keeping and holding down a heavy load is successfully realized. The experimental results show that the performance of the negative admittance control can be arbitrarily adjusted by control parameters and users “load feeling” can be arbitrarily reduced.

Construction of real-time BMI control system based on motor imagery

Recently, a lot of BMIs (Brain-Machine Interfaces) using EEG (electroencephalogram) signals are developed to control external devices such as prostheses and robots. In this paper, in order to develop a new BMI for rehabilitation and/or power support, four different tasks corresponding to different brain excitation degrees are designed. Their EEG spectra are analyzed with short-time FFT, and their features of mu and beta rhythms corresponding to the different tasks are extracted. Finally, one-joint robot arm is controlled by the extracted features, and the proposed approach is confirmed.

Estimation of EMG signal for shoulder joint based on EEG signals for the control of upper-limb power assistance devices

Brain-Machine Interface (BMI) has emerged as a powerful tool for assisting disabled people and for augmenting human performance. Up so far, no studies have succeeded in the power augmentation for the multi-DOFs robot based on EEG signals, especially for the complex shoulder joint. In this work, we propose an electromyography (EMG) estimation method based on electroencephalography (EEG) signals to realize the power assistance. The positions of the electrodes where the motion information of shoulder joint is effectively and exactly extracted are discussed, and a linear model that correlates the EMG to the EEG signal is constructed utilizing motion-related features extracted from multi-location EEG measurements. The constructed model is used to estimate the human muscular activity of shoulder joint from EEG using Principal Component Analysis (PCA) method. The proposed approach is experimentally verified, and an average correlation coefficients are as high as about 0.90 for different subjects are obtained between the estimated and the actually measured EMG signal. Our results suggest that the estimation of EMG based on EEG is feasible. This demonstrates the potential of using EEG signals to support human activities via brain-machine interface.

Development of a light wearable exoskeleton for upper extremity augmentation

This study aims to develop a light and wearable exoskeleton controlled by Electromyography (EMG) signals for upper extremity augmentation to support the elbow and shoulder joints. This exoskeleton has four degrees of freedom, one for elbow joint and three for shoulder joint. The flexion and extension movements of elbow and shoulder joints are actively powered to support the wearer and other two DoFs of shoulder joints are freely passive. The processing of EMG signals and the mechanism and control system are developed. The movements of the exoskeleton are verified.

A novel power add-on unit for attendant propelled wheelchairs with Sensorless Speed Control and Power Assistance

In order to help an attendant who are pushing a conventional commercialized manual wheelchair in which a patient (an elderly or disabled person) is sitting in the wheelchair, in this study, we have been developing a novel power add-on unit (PAU) for attendant propelled wheelchair with sensorless speed control and power assistance. This development is based on the following three concept: to make our PAU (1) easy to be attached to and detached from the wheelchair; (2) able to be used to the most of the conventional commercialized wheelchairs; (3) have a low cost. To enhance its performance, a simple sensorless speed control is experimentally realized. Further, in order to make the attendant more easily and comfortably propel the wheelchair, a sensorless power assist approach is proposed and simulated. The results verified the effectiveness of the proposed two control approaches.