Fall detection for elderly by using an intelligent cane robot based on center of pressure (COP) stability theory

Abstract

An intelligent cane robot was designed for aiding the elderly and handicapped people walking. The robot consists of a stick, a group of sensors and an omni-directional basis driven by three Swedish wheels. Multiple sensors were used to recognize the user's “walking intention”, which is quantitatively described by a new concept called intentional direction (ITD). Based on the guidance of filtered ITD, a novel intention-based admittance motion control (IBAC) scheme was proposed for the cane robot. To detect the fall of user, a detection method based on Dubois possibility theory was proposed using the combined sensor information from force sensors, a laser ranger finder (LRF) and an on-shoe load sensor. The human fall model was represented in a two-dimensional space, where the relative position between the center of pressure (COP) and the center of support triangle was utilized as a significant feature. The effectiveness of proposed fall detection method was also confirmed by experiments.