Identification and optimization of the operator’s hand and a haptic device dynamic, using artificial intelligence methods

A haptic device (HD) is used to simulate a virtual object (VO) for its operator. Since most real objects can be simulated with a spring and damper, the VO is also implemented as a spring and damper in the discrete domain. On one side, the HD should have a small effective mass and friction so that its dynamics can be neglected in contrast to the dynamics of the VO. On the other side, even the small dynamics of the HD or the operator’s hand affect the stability and performance of the robot. So the dynamic model of the HD and the operator’s hand should be determined accurately. In this research, the mentioned dynamics have been identified on a HD using different artificial intelligence methods of ray optimization method, jaya optimization algorithm, crow search algorithm, Big-Bang Big-Crunch, adaptive dimensional search algorithm, and artificial bee colony optimization (ABCO). To this purpose, the theoretical stability boundary is presented as a function of the HD and the hand’s dynamic parameters, the stiffness and the damping coefficient of the VO, sampling time, and time delay. Simultaneously, with some experiments on the KUKA light weight robot IV (LWRIV), the boundary of experimental stability has been achieved in different situations. Then, using the mentioned methods the values of the HD and the hand’s dynamic parameters are obtained in such a way that the error between the theoretical and the experimental stability boundaries is minimal. Studies show that for predicting the HD parameters all mentioned methods have somehow the same and good accuracy, while the BBBC method has the minimum runtime. For predicting the HD and the operator’s hand dynamic parameters, the BBBC method has the minimum error and runtime.