Simulation modelling for robotics grasping of flexible viscoelastic materials using a discrete element model

A simulation model has been developed for robotics grasping of flexible non-rigid materials predicting deformations occurred by the gripping systems. A new concept for the deformations occurred by the gripping forces has been envisaged to reduce the computational cost on parallel grasping processes and a Discrete Element Model (DEM) has been developed for deformation analysis of 2-D flexible elastic and viscoelastic materials. The proposed model consists of point mass elements in a 2-D grid arrangement, connected to each other by discrete spring and dashpot links, allowing each point mass to have two degrees of freedom. Three different methods have been employed to give numerical solutions for the model assuming a linear elastic system. These methods are backwards substitution, Choleski backwards, and iterative conjugate gradient method. In addition the Newmark-β direct integration method is used to solve the model for a viscoelastic system. A simulation tool has been developed for analysis of different systems and the results may be graphically illustrated. Experiments have been carried out to validate the modelling methods for specific material examples. It is concluded that there is a good comparison between the modelling and experimental results and that DEM gives an acceptable prediction of what could happen in a real case. The model was proved to behave with a quadratic convergence and prediction errors fall within 3% to 4% of the actual values measured for the tested materials (rubber, apple tissue, sponge). The computational cost associated with DEM compares favourably with other methods allowing on-line grasping processes. The work continues to develop a simulation modelling for surgical robotics applications.