Publication

Abstract : Most of all the real world engineering applications are connected with differential equations, which has become complex to compute analytically. Paper discusses the efficient way of computing such partial differential equations (PDE) by means of numerical methods. Comparatively with various emerging numerical methods, finite element method (FEM) makes the task of computation achievable with accurate results irrespective of its drawbacks like computational cost and time complexity. Artificial neural networks (ANN) are the one's combined with FEM to make the work more feasible in solving irregular boundaries and to achieve appropriate results. Write up describes the complete importance of FEM based ANN for solving boundary value problems and shows us a glance of solving a PDE and also explains the computation of inverse problems that are considered to be ill-posed for the coefficient approximation. Finally the paper deals with the applications of FEM based ANN techniques in robotics. Essay explains the importance of these methods in developing kinetic and dynamic models in designing a soft robot for its movement and operation as well as its importance and uses in the practical scenario compared to rigid robots. Finally paper presents the stress and strain analysis of the generated CAD model of a robotic hand in FEM softwares like COMSOL and ANSYS.