The estimation for forward kinematic solution of Stewart platform using the neural network

This paper introduces a kind of the forward kinematic analysis, which finds the 6-DOF motions from a given six cylinder lengths in the Stewart platform. In the case of a parallel manipulator, while the solution of the inverse kinematics can easily be found by the vectors of the links which are composed of the joint coordinates in base and plate frame, the forward kinematic is not easily solved due to the nonlinearity and complexity of the Stewart platform´s dynamic output equation with the multi-solutions. Hence, we introduce a linear estimator using Luenberger´s observer and an estimator using the nonlinear measurement model for the forward kinematic solutions. Since it is difficult to find the parameters of the design for the estimator, we suggest the estimation gain to be learned by a neural network with the structure of multi-perceptrons and the learning method using backpropagation. We show the estimation performance using computer simulations