Approximate Multi-Objective Optimization of Medical Foot Support: Case of 3D Shape Optimization
Splayfoot seems not serious diseases. However, it cause fatigue in the daily life. In that sense, if we can solve these problems it might support daily life much comfortable especially for elder people. To support splayfoot, there are some commercial ones. But they just add small amount of support and not made up for each person. Besides, when the height of support is not suitable for patients, it sometimes makes situation worse. We have tried it and measured by using myoelectric potential measurements, and see differences of three patients. Even if we use the same commercial support its effectiveness differs to each other and sometimes it makes worse. Physical therapists make foot support for each patient but they make them owing to their experience. There are studies on the positions of bone of foot and its portrait of the desired positions has been reported by medical doctors. One of them is called Mizuno standard [1]. In the previous study, we tried to design foot support aiming to make portrait of this standard in vertical space. For that purpose, we used approximate multi-objective optimization using Radial Basis Function network. For validation, we used electromyography again. As a result the foot support by the proposed method showed the maximum reduction in integral of myoelectric. However, in the previous study [2], we only used two design variables, and we have only designed vertical phase. Therefore, we have not completely designed the support. In this study, we use 3D spline expression to make support, and try to design 3D shape of support. In validation, the results of approximate multi-objective optimization show the best reduction in integral of myoelectric, and show the effectiveness of the proposed method.