A New Multi-level Algorithm for Identification and Stochastic Adaptive Control of Industrial Manipulators

摘要

A new multi-level algorithm for error compensating, identification and stochastic adaptive control of industrial manipulators robots is presented. It operates at two levels. The first level is based on hidden Markov models (HMM). An off-line trajectory planning for the trajectography problem is developed. It is an iterative and global algorithm (IGA). We will give various disturbances and errors due to several arguments of the machine and its environment. Reversibility and repeatability errors are studied. Given a bloc of data and observations, the algorithm estimates iteratively the parameters of the robot. These data consists of a serial of horizontal cartesian coordinates. The second level is an adaptive control algorithm. It consists of an on-line stochastic adaptive tracking. It is the ergodic recursive algorithm (ERA). To achieve these two levels, a complete dynamic model is considered. The convergence of the two algorithms is discussed. Control laws for IGA and ERA are given and their stability is proven. Some simulations showing the reliability and the efficiency of our algorithms are given.