A Markerless Human-Robot Interface Using Particle Filter and Kalman Filter for Dual Robots

A more natural means of communicating complex movements to a robot manipulator is where the manipulator copies the movements of human hands. This paper presents a markerless human-robot interface that incorporates Kalman filters (KFs) and particle filters (PFs) to track the posture of human hands. This method allows one operator to control dual robot manipulators by using his/her double hands without any contact devices or markers. The algorithm employs Leap Motion to determine the orientation and the position of the human hands. Although the position and the orientation of the hands can be obtained from the sensor, the measurement errors increase over time due to the noise of the devices and the tracking error. The PFs and KFs are used to estimate the position and the orientation of the human hand. Due to the limitations of the perception and the motor, a human operator cannot accomplish high-precision manipulation without any assistance. An adaptive multispace transformation is employed to assist the operator to improve the accuracy and reliability in determining the posture of the manipulator. The greatest advantage of this method is that the posture of the human hands can be estimated accurately and steadily without any assistant markers. The human-manipulator interface system was experimentally verified in a laboratory, and the results indicate that such an interface can successfully control dual robot manipulators even if the operator is not an expert.