Particle Filter Scheme of Both Shins Gaiting And Abdomen Swing for Motion Estimation And Prediction

抄録

Japan is currently a super-aged society. Gait support is becoming increasingly necessary as a result. In this study, we will develop a particle filter to be incorporated into a 2.5D gait recognition method under development that uses three sensors instead of a camera or other devices. This is because recognition using a camera or other means involves issues such as privacy and the time and effort required to attach the filter. The information used in this recognition method is based on sensor readings of the center points of the torso and legs, and the amount of change is modeled as a system model. The center points of the torso and legs were calculated as an elliptical model and a cylinder model, respectively. In actual simulations using simulation software such as Matlab, we were able to achieve the desired movement of the torso and legs during walking. We believe that this will eliminate errors and disturbances in the model and enable more accurate recognition. In the future, we would like to define an actual model that recognizes the state of walking in this research and develop a walking support machine to support walking. In addition, we plan to further develop the particle filter system model to be able to read special gait such as hemiplegic gait.