As modern society becomes increasingly more elderly, developing specialized equipment to help assist this growing population in everyday tasks becomes more and more important. Here we have developed a passive gravity balanced arm support system called the "Arm-Balancer". It was designed to assist the arm motion of elderly people with disabilities. The objective of this study is to determine the utility of the "Arm-Balancer". We will evaluate the effect of this device on the arm muscle tension of the user. The "Arm-Balancer" is supported by a gas spring to eliminate gravity and may be attached to a chair, a bed or used on the floor when attached to the "Tatami Mat" seat. The "Arm-Balancer" consists of a two-segment exoskeleton designed for the users upper arm and forearm. The links are made from stainless steel rods whose lengths are designed to parallel the lengths of the users upper arm and forearm. The upper arm segment is telescopic so the length can be adjusted to the size of the users. Driving the evaluation we determined that the "Arm Balancer" had an assistive force of 10N to the arm and 5N to the wrist. Some elderly people with disabilities were receptive to the "Arm Balancer" while others were ambivalent.

View full abstract

Electromyogram (EMG) is interference waveform of motor unit action potentials (MUAPs) within the detection area. We assumed that value obtained by detecting and counting peaks of the EMG corresponds to number of MUAP included in the EMG. In the previous study, it was known that increase of peak number becomes smaller and the number becomes saturated as muscle contraction level becomes higher. Therefore, we modeled and formulated relationship between numbers of peaks and MUAPs in the EMG and proposed an estimation method using them. For this purpose, the method was applied to simulated and real surface EMGs and the utility was discussed. As the result, in the simulated EMG at eight muscle contraction levels, the estimated value approximated the actual number and the error decreased from 37% to 11% at the maximum contraction levels. In the real surface EMG, the relationship between muscle contraction level and estimated value showed the same tendency as the simulated EMG. Therefore, it was suggested that possibility of estimating the number of MUAP in the EMG by using the method.

View full abstract

Mat sensors are used to detect movement of a person who is unable to walk and requires assistance when getting out of bed to keep from falling. Mat sensors are also used to detect a person who might wander the facility unattended. We intoroduced a system which detects the action of getting out of bed using an ultrasonic array sensor to monitor head position. Our previous paper showed the detection rate is 84.6% with an error rate 5.4% and a loss rate of 10.0% for getting out of bed. Though there was only a subject and a bed in this experimental environment, there are various objects and other persons around the actual bed. In actuality, there are objects such as a drip stand and a chair or person, such as a nurse or attendant, around the bed. So, we needed to consider the unwanted ultrasonic reflection from them that might lower the detection rates above. We then created a test environment with products and the person mentioned above around the bed. We used 7 people in 2 test situations in which each person got out of bed 17 differrent ways. We have improved our detection

View full abstract

This paper proposes a compensated-norm-ratio Φ_N⁺ algorithm to improve the separation of a submerged pulse wave in a heavy tapping artifact that is measured by two wave lengths (denoted by IR, R ) of SpO₂. This algorithm employs a 2×2 inverse matrix to separate the pulse wave. Parameters Φ_S and Φ_N are unknown for the inverse matrix. Φ_S and Φ_N are respectively the R/IR norm ratio of the pulse wave and the artifact wave for the observed signals. It is able to determineΦ Φ_S from the stable pulse wave segment. By contrast,careful estimation is necessary for Φ_N. This is because, _N is shifted from Φ_N^true due to the pulse wave superimposed on the artifact segment. This shift induces poor pulse wave separation. A compensation algorithm is necessary to make Φ_N closer to Φ_N^true. The proposed compensated-norm-ratio Φ_N⁺ results in better separation of the pulse wave than that obtained using Φ_N. This algorithm is also effective for case with other body movement artifact. This algorithm can not only be not be applied to two-wave pulse photometry but can be also extended to multi-wave pulse photometry.

View full abstract

Tripping over which is one of the causes of falling is caused by luck of recognizing obstacles. The 3D motion of five young adults was analyzed for stepping over an obstacle of 48mm high. Subjects who were more likely to trip over in usual activities had longer minimum distance between toe and obstacle. This result meant subjects who were more likely to trip over stepped over the obstacle more safety and contradicted the result of sensory test asked whether subjects were likely to trip over in usual activities. This suggested that one of the causes of tripping over was luck of recognizing obstacles in case of people who were more likely to trip over in usual activities. The effect of the visual sensation on human motion was also studied by using of eye tracking. Eye motion of five young adults was captured by eye tracking system during walking simulation of walking toward an obstacle using screen image.Subjects who were more likely to trip over were less likely to gaze at an obstacle. This result supported the results of motion analysis. Those results suggested that one of the main causes of tripping over was luck of recognizing obstacles and that it was possible to extract the effect of recognition on human motion by using of motion analysis and eye tracking.

View full abstract

In order to design a prosthesis, it is important to consider fitting between skin of the limb and the socket. The purpose of this study is to quantify the normal force between the skin of the lower limbs and an inner socket in the lower limb prosthesis. To measure normal pressure sensors were set on the inside of prosthesis socket. Pressure and stress were collected during static standing, stepping and running motions. Changes in the prostheses, inner sockets, and alignment were investigated based on the contact forces. It was found that the distribution of normal stress changed according to the motion: maintaining static posture and dynamical motion such as walking. The relation between slipping of the stamp and the normal stress in the static posture diddered from that in running.

View full abstract