HUMAN GAIT BEHAVIOR INTERPRETATION BY A MOBILE HOME HEALTHCARE ROBOT

The ultimate goal of this study is to develop autonomous mobile home healthcare robots which closely monitor and evaluate the patients' motor function, and their at-home training therapy process, providing automatically calling for medical personnel in emergency situations. The robots to be developed will bring about cost-effective, safe and easier at-home rehabilitation to most motor-function impaired patients (MIPs), and meanwhile, relieve therapists from great burden in canonical rehabilitation. In order to achieve this goal, we have developed the following programs/algorithms for monitoring human activities and recognizing human behaviors: (1) control programs for a mobile robot to track and follow a human subject by three different viewpoints; (2) algorithms for analyzing lower limb joint angles from RGB-D images from a Kinect sensor setup at a mobile robot; and (3) algorithms for recognizing human gait behavior. In (1), side viewpoint, front/back viewpoint and a middle angle viewpoint (between two former viewpoints) tracking were developed. In (2), depth image compensation with colored markers was implemented to deal with the skeleton point extraction error caused by mixing-up and frame flying of depth image during tracking and following human subjects by the mobile robot. In (3), we have proposed a hidden Markov model (HMM) based human behavior recognition using lower limb joint angles and trunk angle. Experimental results showed that joint trajectory could be measured and analyzed with high accuracy compared to a motion tracking system, and human behavior could be recognized from the joint trajectory.