Robot Foot Trajectory Tracking Motion Control Using Kinect

As for the actions of the human, the legs action behaves are more complex articulations and are more easily affected by the number of degrees of freedom and kinematics modeling errors. In this paper, the robot foot trajectory tracking motion control using Kinect sensor is proposed. Kinect sensor is used to build a robust 3D legs gesture recognition system to achieve operator's legs action tracking such that the legs of the robot in the remote site can duplicate that of the operator. It is thus a very attractive and challenging problem to transfer the leg gestures of the human after recognition to that of the robot. The robust leg gestures recognition is proposed to handle the legs kinematics and inverse kinematics from the legs' ankle trajectory obtained from the Kinect. According to the legs inverse kinematics, the distance metric using the ankle tip parts of the operator is used to measure the dissimilarity of the operator's legs shapes and ankle trajectory. The merits by using ankle tip parts are that it can easily distinguish the leg gestures by using inverse kinematics between of the operator and robot even if the degrees of freedom between them are different or less. The experiments demonstrate that the proposed leg gesture recognition system is accurate with 95% mean accuracy on a challenging 5-gesture dataset, distortions, orientation, or scale changes, and can work in uncontrolled environments in cluttered backgrounds.