Abstract

The distribution of ground reaction forces (GRFs) is important information for the precise diagnosis of diseases and the calculation of gait-related kinetic quantities, such as the center of pressure, abnormalities, and so on. However, many insole-type mobile GRF sensing systems measure only discrete GRFs because of a limited number of sensors. In this paper, a kernel density function is proposed to estimate the force distribution of the foot from measurement with the limited number of sensors. For precise verification, the foot pressure during walking is measured by force plates mounted on a laboratory floor and by four pressure sensors embedded in an insole. The sensor measurements themselves cannot represent the total foot pressure compared to the force plate due to the limited sensing area; however, with the proposed method, the pressure of the unmeasurable area of the insole can be estimated. The parameters of kernel density function are derived based on the anatomic information of the human foot. The experimental result shows that the foot pressure of unmeasurable area is successfully estimated by the proposed method, and the insole system can measure the total foot pressure even with a limited number of sensors.