The purpose of FES is to control the movements of the human body by applying electrical stimulation. Early studies used simple open-loop control for practical application, with little feedback control, so I wanted to achieve more advanced engineering control. With feedback control, measurement of movements is also important, so I gradually shifted the focus of our research to develop a measurement method. Eventually, as I learned more about the complexity of the mechanisms inherent in humans, I came to realize that engineering system alone was not enough to achieve the precise motor control. At the same time, I also felt that the global trend in FES research was gradually shifting from motor control to applications in motor rehabilitation. So, as I moved the focus of my research to rehabilitation, I began to think about whether we could realize a system that would simplify control of movement and apply it to rehabilitation. Without a proper understanding of the patient’s condition, it would not be possible to apply the system to the rehabilitation program, so I also conducted a study on measuring and evaluating the patient’s motor function. Since joining the Graduate School of Biomedical Engineering, I have been developing a system to evaluate movements and a system to assist patients’ movements in their rehabilitation training. Recently, I went back to the beginning and resumed my research on how to control complex movements in an engineering manner.
When controlling human movements by FES, it is necessary to solve the problems of nonlinearity, time variability, redundancy, and individual differences in humans who are controlled. We would like to solve all of these problems at once, but it’s not easy. What we are focusing on now are redundancy and individual differences. For the redundancy, we would like to improve it to a point where it is practically viable. Regarding individual differences, we are working on a method for learning the characteristics of a target person using a neural network. We are also developing a method of feedback control, in which a sensor measures the current state of the subject, predicts the subject’s movement in a few seconds, and sets a target for the movement control. After all, the movement of a person is low in reproducibility, differs from time to time, and it is necessary to control it according to the person’s condition. In addition, I think it would be useful for rehabilitation if we could embed in the system a mechanism that imposes a small load on the patient’s movements, or that can change the patient’s gait for the better with a little help.
In the research on the measurement of movements, we use inertial sensors, mainly accelerometers, and gyroscopes. Inertial sensors are convenient because they are easy to put on and off, are lightweight, don’t interfere with movement, and can measure the condition of the body segments they are attached to. In our current studies, we have a total of seven sensors on the lumbar region (trunk), thighs, shanks, and feet (on shoes) of both lower limbs. In the case of a healthy person, it can be measured with relatively good accuracy. For example, the angles of lower limbs in the sagittal plane during walking can be measured with an error of about several degrees on average. Stride length and walking speed can also be estimated with an average error of about 2-3%. Inertial sensors are not good at measuring intense movements, but they can be sufficiently applicable to relatively slow movements such as those of motor disabled patients. On the other hand, measurements with an optical motion analysis system are accurate, but they can only be made in limited situations, and patients also work hard unconsciously when they are caught on camera. That is why I think the inertial sensor can measure the patient’s natural condition and also properly measure the patient’s condition after returning home. Another advantage of the inertial sensor is that it can easily measure anytime and anywhere. Some hospitals don’t have enough space to take measurements, so we would like to develop a system that can be used easily in such places. In the medical field, including rehabilitation, I think it is possible to separate the initial screening phase into simple measurements with this kind of device, and then a more detailed diagnosis is performed with a high-precision device.
In our research on rehabilitation, we are developing a system that combines measurement and electrical stimulation to practice walking by giving electrical stimulation while measuring with sensors. Specifically, we use electrical stimulation to correct the drop foot (a condition in which the ankle cannot be made dorsiflexion, so the patient cannot walk without swinging the lower limb outward or lifting the foot high during walking). We are trying to electrically stimulate the tibialis anterior muscle and the common peroneal nerve during the swing phase and measure how the gait changes and how it changes between wearing a brace such as ankle-foot orthosis (AFO) and correcting the drop foot with FES. Of course, some patients find the orthosis more stable and easier to walk when walking long distances. However, in Japan, there is a culture of taking off shoes and sitting on the tatami at home, so I hope to provide an environment that allows people to adapt to this kind of lifestyle without using an orthosis, and that is what we are researching.