Neural Network and Kinematic Control of a Stewart Platform for Animatronic Application

This paper presents the design and development of an animatronic robot neck based on a Stewart platform. Trajectory tracking control of the head is achieved through inverse kinematic relations developed analytically and using a Deep Neural Network (DNN) approach. Both approaches are validated using MSC ADAMS ${ }^{\text {TM }}$. Additionally, the same is utilized to determine torque requirements, eliminating the need for traditional analytical dynamics modeling. An Arduino Due development board is employed to execute control algorithms using the Simulink support package for Arduino hardware within the MATLAB-SIMULINK environment. The successful physical realization of the system on the Arduino platform demonstrates the feasibility and effectiveness of the proposed approach, reaffirming its practicality and potential for further development. Results indicate that the neural network model performs as effectively as the analytical model in accurately predicting the servo arm angular displacements for given platform positions.