A new possible branch of computational intelligence for developing real time control in mechatronics

Recent advances in developing a new branch of soft computing (SC) developed for the control of mechanical devices is reported. Instead of using the architectures of traditional SC it is based on uniform structures obtained from the mathematical framework of classical mechanics. In connection with that the concepts of "minimum operation symplectizing algorithm" (MOSA) and "situation-dependent partial system identification", were developed. The advantages are as follows: (1) a priori known and limited number of parameters; (2) parameter tuning is replaced by simple, explicit, finite, and standard algebraic computations; and (3) to refine the "core algorithm" several interpolation techniques can be incorporated in it. To illustrate its operation simulation examples are given for a special 3-DOF robot arm under a significant environmental interaction.